Seroprevalence of Toxoplasma gondii infection and associated risk factors in Huicholes in Mexico

Background Very little is known about the seroepidemiology of Toxoplasma gondii infection in ethnic groups in Mexico. Huicholes are an indigenous ethnic group living in a remote mountainous region in Mexico. We sought to determine the prevalence of anti-Toxoplasma IgG and IgM antibodies in Huicholes; and to determine the association of Toxoplasma seropositivity with socio-demographic, behavioral, and clinical characteristics of Huicholes. Methods We performed a cross sectional survey in Huicholes from September 2013 to January 2014. A convenience sampling method was used. We investigated the prevalence of anti-Toxoplasma IgG and IgM antibodies in 214 Huicholes using enzyme-linked immunoassays. A standardized questionnaire was used to obtain the characteristics of the Huicholes. Bivariate and multivariate analyses were used to assess the association of Toxoplasma exposure and Huicholes’ characteristics. Results Of the 214 Huicholes studied (mean age: 37.98 ± 15.80 years), 71 (33.2%) were positive for anti-T. gondii IgG antibodies and 47 (66.2%) of them were also positive for anti-T. gondii IgM antibodies. Seroprevalence of T. gondii infection did not vary with age, sex, or occupation. However, seroprevalence of anti-T. gondii IgM antibodies was significantly higher in female than in male Huicholes. Multivariate analysis of socio-demographic and behavioral characteristics showed that T. gondii exposure was associated with consumption of turkey meat (OR = 2.28; 95% CI: 1.16-4.46; P = 0.01). In addition, seroprevalence of T. gondii infection was significantly higher in Huicholes suffering from dizziness and memory impairment than those without such clinical characteristics. Conclusions Our results demonstrate serological evidence of T. gondii exposure among Huicholes which may be impacting their health. Results of this first study of T. gondii infection in Huicholes may be useful for the design of optimal preventive measures against infection with T. gondii.


Background
Infections with the parasite Toxoplasma gondii (T. gondii) occur worldwide [1] and affect about one third of humanity [2]. Although most infections with T. gondii are asymptomatic, some infected individuals may suffer from symptomatic pathological changes in the lymph nodes, eyes, and central nervous system [3]. In addition, pregnant women with primary infection with T. gondii may transmit the infection to the fetus leading to congenital disease [4]. Immunocompromised individuals infected with T. gondii may develop a life-threatening disease [5]. Ingestion of food or water contaminated with oocysts shed by cats [3,6] and eating undercooked or raw meat containing tissue cysts [3,7,8] are important routes of T. gondii transmission.
Very little is known about the epidemiology of T. gondii infection in ethnic groups in Mexico. We have previously studied the seroepidemiology of T. gondii infection in Mennonites [9] and Tepehuanos [10] in Durango, Mexico. To the best of our knowledge, there is not any report about the epidemiology of T. gondii infection in Huicholes (an indigenous ethnic group living in a remote mountainous region (Sierra Madre Occidental) in the western central Mexican states of Nayarit, Durango, Jalisco and Zacatecas. Life style in Huicholes differs from that in other rural population groups in Durango; they live in marked poverty with very poor housing and sanitary conditions. They have limited access to health care services, and Hospitals in their region do not have a number of laboratory tests for diagnosis of infectious diseases i.e., infection with T. gondii. It is important to study the epidemiology of T. gondii infection in Huicholes because they live in a climatic scenario that may favor T. gondii infection. Huicholes live in a warmer and more humid area than other population groups in the region. Environmental factors may contribute to a higher seroprevalence of T. gondii infection [1,11]. In general, the seroprevalence of T. gondii infection is higher in humid climates than in dry climates; and this is the case in humans [12][13][14] and animals [15][16][17]. Furthermore, Huicholes eat meat from wild animals that may be infected with T. gondii. Therefore, we sought to determine the seroprevalence of T. gondii exposure in Huicholes and the association of Toxoplasma seropositivity with socio-demographic, behavioral, and clinical characteristics of Huicholes.

Study design and study population
We performed a cross sectional survey in Huicholes in Mexico from September 2013 to January 2014. Huicholes were sampled in the locality of Huazamota in the municipality of El Mezquital in Durango State, Mexico. Huazamota (23°28´N 104°24´W) has an altitude of 600 meters above sea level, a warm-sub-humid climate, and a mean annual temperature of 19.2°C. The south region of El Mezquital municipality has a mean annual rainfall varying from 800 to 1000 mm. Other ethnic groups live in the mountainous region including Mexicaneros and Tepehuanos. Inclusion criteria for the study subjects were: 1) Huichol ethnicity (people who speak the Huichol language and identify themselves as Huicholes); 2) aged 14 years and older; and 3) that voluntarily accepted to participate.

Sample size and sampling method
For calculation of the sample size, we used a reference seroprevalence of 22.4% [10] as expected frequency of the factor under study, 7,000 as the size of population from which the sample was selected, 16.9% as the least acceptable result, and a confidence level of 95%. The result of the calculation was 214 subjects. A convenience sampling method was used. Specifically, the authors approached Huicholes leaders for permission and support; each leader communicated and invited all people under his command; those who accepted the invitation gathered in a specific area to provide socio-demographic data and blood sample; 214 people who met the inclusion criteria were enrolled.

Socio-demographic, clinical, and behavioral data
We used a standardized questionnaire to obtain the sociodemographic, clinical and behavioral characteristics of the Huicholes. Socio-demographic items were age, sex, birth place, residence, educational level, occupation, and socioeconomic status. Clinical data included the presence of underlying diseases, presence or history of lymphadenopathy, frequent presence of headache, dizziness, impairments of memory, reflexes, hearing, and vision, and a history of surgery, blood transfusion or transplants. Clinical data including impairments was self-reported. Huicholes were considered "ill" when they suffered from any disease either acute or chronic affecting any organ or system i.e. digestive, respiratory, circulatory, endocrine, or nervous, and included any psychiatric, rheumatic, hematological or nutritional disorder and any type of morbidity. In Huichol women, obstetric data were also obtained. Behavioral items included animal contacts, contact with cat excrement, foreign travel, meat consumption (pork, beef, goat, lamb, boar, chicken, turkey, pigeon, duck, rabbit, venison, squirrel, horse, opossum, or other), frequency of meat consumption, consumption of raw or undercooked meat, unpasteurized milk, dried or processed meat (ham, sausages or chorizo), consumption of unwashed raw vegetables, fruits, or untreated water, frequency of eating away from home (in restaurants or fast food outlets), contact with soil (gardening or agriculture), and type of flooring at home from all participants were obtained.
Serological examination for T. gondii antibodies Serum samples were obtained from about 3 ml of whole blood. Sera were kept frozen at −20°C until analyzed. All sera were analyzed by qualitative and quantitative methods for anti-T. gondii IgG antibodies with a commercially available enzyme immunoassay "Toxoplasma IgG" (Diagnostic Automation Inc., Calabasas, CA, USA). Anti-T. gondii IgG antibody levels were expressed as International Units (IU)/ ml, and a cut-off of ≥ 8 IU/ml was used for seropositivity. In addition, sera positive for T. gondii IgG were further analyzed for anti-T. gondii IgM antibodies by a commercially available enzyme immunoassay "Toxoplasma IgM" kit (Diagnostic Automation Inc., Calabasas, CA, USA). The cut-off for IgM seropositivity for each assay was obtained by multiplying the mean cut-off calibrator optical density by a correction factor (f = 0.35-0.40) printed on the label of calibrator. All assays were performed following the instructions of the manufacturer and included positive and negative controls in each run. A positive IgG test and a negative IgM test in a participant was interpreted as a latent infection. A positive IgG test and a positive IgM test in a participant was interpreted as probability of a recent or acute infection.

Statistical analysis
We used the Epi Info version 3.5.4 and SPSS version 15.0 software for the statistical analysis. The Pearson's chisquare test and the Fisher exact test (when values were less than 5) were used for initial comparison of the frequencies among groups. Multivariate analysis was used to assess the association between the characteristics of the Huicholes and the seropositivity to T. gondii. Variables were included in the multivariate analysis if they had a P value equal to or less than 0.15 in the bivariate analysis. Odds ratio (OR) and 95% confidence interval (CI) were calculated by multivariate analysis using the Enter method. The Hosmer-Lemeshow goodness of fit test was used to assess the fitness of the regression model. A P value <0.05 was considered statistically significant.

Ethical aspects
The purpose and procedures of the survey were explained to all Huicholes. This study was approved by the Ethical Committee of the General Hospital of the Secretary of Health in Durango City, Mexico. Participation in the study was voluntary and a written informed consent was obtained from all participants and from the next of kin of minor participants. All Huicholes were proficient in Spanish and understood explanations about the purpose and procedure of the survey as well as the informed consent provided by the interviewers. Results of the laboratory tests were sent to the Huicholes' nearest Hospital (Huazamota) where health care providers could inform participants about their results and provide them with medical care if needed.

Results
In total, we enrolled 214 Huicholes in the study including 86 (40.2%) males and 128 (59.8%) females. Most Huicholes were born in Durango; their mean age was 37.98 ± 15.80 years (range 14-82 years). General sociodemographic characteristics of the 214 Huicholes studied are shown in Table 1.
Concerning clinical characteristics, seroprevalence of anti-T. gondii IgG was significantly higher in Huicholes suffering from dizziness and memory impairment than those without such clinical characteristics ( Table 2). The frequency of T. gondii exposure in subjects with dizziness and memory impairment did not vary with age (P = 0.19 and P = 0.48, respectively). The frequencies of other clinical characteristics including the presence of underlying diseases; suffering from frequent headaches; presence or history of lymphadenopathy; reflexes, hearing and visual impairments; histories of surgery, blood transfusion, and transplant were similar among T. gondii positive and T. gondii negative individuals. Histories of miscarriage and stillbirth in women were not associated with T. gondii seropositivity. With respect to behavioral characteristics, a number of variables showed P values lower than 0.15 in the bivariate analysis including presence of dogs at home (P = 0.06), consumption of raw dried meat (P = 0.14), and consumption of meat from goat (P = 0.14), turkey (P = 0.002), and pigeon (P = 0.04). A selection of behavioral characteristics and their correlation with T. gondii exposure are shown in Table 3. Other behavioral characteristics of Huicholes including contact with cats, cleaning cat excrement, raising animals, traveling, consumption of meat other than goat, turkey and pigeon meat, frequency of meat consumption, degree of meat cooking, consumption of unpasteurized milk, processed meat, unwashed raw vegetables or fruits, untreated water, frequency of eating out of home, soil contact, and soil flooring at home showed P values higher than 0.15 in the bivariate analysis. Further analysis by using logistic regression showed that T. gondii exposure was only associated with consumption of turkey meat (OR = 2.28; 95% CI: 1.16-4.46; P = 0.01) ( Table 4). A P = 0.35 was obtained in the Hosmer-Lemeshow test indicating an acceptable fit of our regression model.

Discussion
The present study was performed to investigate the seroepidemiology of T. gondii infection in Huicholes in Mexico. Results indicate that Huicholes have one of the highest seroprevalences of T. gondii infection reported in the region. The seroprevalence found in Huicholes (33.2%) is  higher than the mean (23.8%) seroprevalence of T. gondii infection reported in the general population in rural areas in Durango State [18]. In addition, the seroprevalence in Huicholes is higher than the 6.1% seroprevalence of T. gondii infection reported in urban general population in the capital Durango City [19]. With respect to other ethnic groups in the region, the seroprevalence found in Huicholes is comparable with the 30.3% seroprevalence of T. gondii infection reported in Mennonites [9] but is higher than the 22.4% seroprevalence reported in Tepehuanos [10]. Huicholes and Tepehuanos live in the same mountains (Sierra Madre Occidental), however, Huicholes live in more remote places deeper into the mountainous region than Tepehuanos. It is known that the seroprevalence of T. gondii infection varies depending on the climate conditions in the communities, i.e., a low seroprevalence in dry and hot climate [14], and high seroprevalence in humid regions [12]. Huicholes communities are located in a lower region on the mountains with warmer and more humid climate than the one of the Tepehuanos settlements. However, difference in the seroprevalences among Huicholes and Tepehuanos should be interpreted with care because of an age limitation of the comparison: the mean age in Tepehuanos (31.03 ± 16.71 years old) was lower than the one (37.98 ± 15.80 years old) in Huicholes. Toxoplasma exposure has been linked to low socioeconomic status [20,21], and such characteristics might have contributed to the increased seroprevalence of T. gondii infection in Huicholes. The seroprevalence of T. gondii infection in Huicholes with low socioeconomic status (34.3%) was higher than in those with medium socioeconomic status (15.4%). However, such increase was not statistically significant because of the limited number (n = 12) of subjects with medium socioeconomic status in the comparison. In the present study women outnumber men. The higher number of women than men in the study may be due to a number of factors including refusal of some men to participate, and more migration and violent deaths in men than in women in the study region. However, the imbalance in sexes in this study is unlikely to affect the seroprevalence rate since seroprevalence of T. gondii infection has been found similar in men and in women of general populations in rural [18] and urban [19] Durango, Mexico.  Multivariate analysis of the socio-demographic and behavioral characteristics of the Huicholes allowed us to identify that consumption of turkey meat was positively associated with T. gondii exposure in Huicholes. In a previous study in the general population in rural Durango, consumption of turkey meat was also associated with T. gondii exposure [18]. In addition, in a study in pregnant women in the urban capital Durango City, consumption of turkey meat was associated with T. gondii infection too [22]. Turkey meat is a potential source for infection with T. gondii [23]. Experimental T. gondii oocyst infections in turkeys have shown the parasite spreading over the whole organism as determined by polymerase chain reaction [24]. In a previous study in birds in Durango, no serological evidence of T. gondii infection in 16 turkeys (Meleagris gallopavo) was found [25]. However, the number of studied turkeys was too small to exclude T. gondii infections in turkeys in Durango. Turkey meat is frequently cooked in big pieces; therefore, it is likely that this meat remains undercooked in some deep areas.
Seroprevalence of T. gondii infection usually increases with age in our region [18,19,26]. In the present survey, seroprevalence tended to increase with age; however, such increase did not reach statistical significance (P = 0.09). It is likely that the small sample size of the oldest Huicholes subgroup prevented us obtaining a statistically significant difference.
Remarkably, in the present study we found an association of T. gondii exposure with the presence of dizziness and memory impairment in Huicholes. This finding may indicate a causal association of infection with T. gondii and central nervous system illness in Huicholes. The association of T. gondii exposure with dizziness and memory impairment was also found in a recent study in migrant agricultural workers in Durango [21]. The association of memory impairment and T. gondii exposure was also assessed in other ethnic groups in Durango including Mennonites [9] and Tepehuanos [10]; however, no association was found. The association of dizziness and T. gondii exposure was not assessed in Mennonites [9] and Tepehuanos [10]. We are not aware of further reports on the association of dizziness with T. gondii infection. On the other hand, the association of memory impairment with T. gondii infection found in the present study agrees with previous reports [27,28]. In a previous study in gardeners in Durango City, T. gondii seropositivity was associated with memory impairment [27]. In a recent study in seniors in Germany, researchers found that T. gondii seropositivity was associated with a reduction of about 35% in working memory, a lower performance in verbal memory, and a decreased quality of life [28]. A number of reports indicate that T. gondii infection may lead to neurological and behavioral changes. Experiments in adult mice have shown that infections with T. gondii cause neurological and behavioral abnormalities secondary to inflammation and loss of brain parenchyma [29]. In addition, chronic infections with T. gondii in mice can damage the spatial learning and memory capability [30]. The behavioral alterations associated with T. gondii infection in humans and animals were recently reviewed by Flegr [31] and Webster et al., [32]. Intriguingly, latent T. gondii infection was associated with improvements in cognitive control processes in young healthy humans [33]. However, this effect might perhaps not be found or even reversed in old age. Beste et al. [34] found that latent T. gondii infection leads to deficits in goal-directed behavior in otherwise healthy elderly individuals.
Anti-T. gondii IgM antibodies were present in a high number of anti-T. gondii IgG positive Huicholes. Such high frequency of IgM positive results was unexpected and should be interpreted with care since IgM ELISA kits have a high rate of false positive results [35]. Therefore, discrimination between recent and latent T. gondii infections is not accurately obtained by ELISA. Other methods such as IgG avidity [36,37] may aid in such discrimination.
The present study has some limitations, including the sampling method and the small sample size of elderly Huicholes. We were unable to perform random sampling because participation of Huicholes in the study depended to a large extent on the permission of the Huicholes leaders. Huicholes leaders asked their people to participate and only Huicholes who attended the invitation were sampled. A small number of elders attended the invitation.

Conclusion
Our results demonstrate serological evidence of T. gondii exposure in Huicholes and Toxoplasma may be impacting their health. This is the first report of T. gondii infection in Huicholes, and results should be useful for the optimal design of preventive measures against T. gondii infection.