Socioeconomic Determinants of Chagas Disease in an Endemic Zone in Western Mexico

Background and Objectives: Chagas disease (ChD) is traditionally associated with poverty and ignorance. However, we do not have hard evidence about this phenomenon. Our objective was to estimate ChD’s seroprevalence in western Mexico, analyzing its association with socio-demographic variables. Methods : We surveyed 1264 habitants of 58 rural and urban localities in Jalisco, Colima, and Nayarit, Mexico, identifying the ChD through indirect hemagglutination and Western Blot tests. We recorded socio-demographic and house characteristics, analyzing their association with ChD using Multiple Logistic Regression. Results : We found 85 ChD cases (seroprevalence of 6.7%). We found a significant correlation with low SES young females in rural areas (OR= 3.28; p= 0.002). However, no significant association of ChD with the SES, the SL, or the house’s quality was noticed. Conclusion : Young women living in rural areas with a low SES represent a risk population that deserves attention for preventive programs. We did not find an association of ChD with the rest of the variables in persons older than 31 years, suggesting that ChD transmission in this region is not a predominantly poverty or ignorance problem.


Introduction
Chagas Disease (ChD) is an originally sylvatic parasitic infection caused by the hemoflagellated protozoan Trypanosoma cruzi (Tc). The parasite is transmitted by hematophagous insects of the subfamily Triatominae (Hemiptera, Reduviidae), constituting a Nested infection that involves diverse wild mammals [1,2] and eventually humans, in whom the parasite produces the acute or chronic forms of ChD. Identifying those human populations at risk of infection allows prioritizing strategies to control or prevent ChD [3]. Many authors have postulated that transmission to humans occurs mainly in rural settlements with deprived housing in impoverished regions, usually in huts made of adobe, with dirt floors or palm roofs, surrounded by corrals of domestic animals, resembling its original nidus [4,5]. However, hard evidence concerning the association between poverty, marginalization, low education, and inadequate housing with ChD prevalence is scarce and not directly correlated to ChD, but only to Triatomine infestation.
In Mexico the average prevalence of ChD is estimated around 3.3% [6], especially in states of the country considered the poorest, such as Chiapas, Veracruz, Oaxaca, and Guerrero, where we found the highest ChD rates [7]. However, in the western Pacif-ic coast, with higher SES and SL, many authors have documented ChD in urban communities [8,9]. To analyze specifically the association between ChD prevalence with the SES, SL, Sex, Age, housing and labor conditions, we carried out a serosurvey in rural and urban settlements in Jalisco, Nayarit, and Colima in the western region of Mexico.

Target population and study design
We performed a population based cross-sectional seroepidemiological survey among Jalisco, Colima, and Nayarit inhabitants in western Mexico (see figure 1) from September 2010 to May 2011. The studied area comprises approximately 112000 km2, with 9 723 667 inhabitants [10]. In total, we selected 58 urban and rural localities in the region, sampling 1264 subjects. We applied a structured questionnaire and took a venous blood sample to identify anti-Tc antibodies using the Indirect Hemagglutination (IHA) and the Western Blot (WB) test in approximately twenty people randomly selected in each locality from an average of ten houses. The inclusion criteria were: age from 1 to 85 years; with all their lifespan or approximately ten years of permanent residence in the locality (to link a prolonged exposure time with environmental factors in the community, and to minimize the possible effect of carrying Tc from other places). We excluded patients with severe chronic disease or contraindication for venipuncture.

Laboratory test
We separated serum from blood samples and stored at -70 °C until their process. Then, our team performed IHA with the commercial kit (Chagatest, Wiener Labs ® ) following the manufacturer instructions treating the samples with 2-mercaptoethanol (2-ME) dilution of 1:40, considering a specimen positive when agglutination occurred at a dilution 1:40 or higher. The positive IHA samples and a stock of negative cases were analyzed using the WB test with a standard immunoelectrotransference technique [11] in the National Institute of Cardiology Ignacio Chavez in Mexico City. We considered ChD's cases those with two positive tests, expressing this in a nominal dichotomous scale.

Socioeconomic variables Socioeconomic Status (SES):
We considered four categories according to the Mexican Association of Opinion and Marketing [12], this includes the occupation of the head of the family; access to urban services; quality of housing; spending on education, leisure and health. The categories were: 1. High SES: People with skilled jobs, complete houses, public services, and access to private medical care or insurance; 2. Medium SES: Those persons with a formal job and beneficiaries of public social security, but with modest housing and without automobile; 3. Poverty level: Family´s head dedicated to informal commerce and non-qualified or elementary jobs, with poor housing conditions and more than three people per room; and finally 4. Extreme poverty: People living in incomplete houses, with meager income, many of them working as peasants or unemployed. In each family, we included all the participant members under the same category.

School Level (SL):
We considered it only in older than 11 years and stratified them in four categories: 1. Without schooling (individuals with no or incomplete primary education); 2. High school, or early technical education; 3. Technical or administrative careers; and 4. People with higher education as professionals or postgraduate.
Occupation: We stratified this variable in people older than 18 years in six different groups according to the International Standard Classification of Occupations (ISCO) [13]: a. Field workers, including farmers, peasants, fishers; b. Services and Sales, like store clerks, street vendors, musicians, waiters, etcetera; c. Managers and professionals; c. Workers in Factories, masons, mechanics, craftsman; d. Students; and e. House workers including homemakers and maids Type of locality: The categories were: 1. Rural towns with less than 2500 inhabitants, and 2: Urban localities with more than 2600 inhabitants (up to 820,000) according to the INEGI11.

Animals at home:
We considered the presence of pets, farmyard or stable animals (poultry, pigs, horses, cows, goats) inside house or in backyard.
House characteristics: We stratify it according to the construction material and the house's finishes that could favor the infestation of triatomines. For this, we considered the materials of Ceiling, Walls, and Floor, being the worse those with a dirt floor, no plastered walls or roof made of palm or cardboard, and the best, houses made of plastered bricks, ceramic floor and house commodities like conditioned air or internet connection.
Inhabitants per room-Two or less people per room, or more than three persons per room.

Statistical analysis
We carried out the correlation analysis of ChD seroprevalence with gender, age, SL, SES, house characteristics, animal ownership, type of settlement, and occupation using Multivariate Logistic Regression, with estimation of Odds Ratio (OR) and their Confidence Intervals to 95% (CI95%) with the Woolf's method.
Ethical aspects: The institutional committees for Ethics and Research of the Secretary of Health in Colima, Jalisco and Nayarit approved the protocol (#142/09). We obtained written informed consent from all participants (parents or legal guardians in the case of minors) before administration of the questionnaire or blood sampling.

Results
We obtained 1264 valid samples, of which 85 resulted positive to ChD, representing a prevalence of 6.72% (CI95%: 5.4 to 8.25). Table 1 shows the total individuals' social and demographic conditioning factors and their entailment with Tc infection. We sampled 371 men in whom 18 were positive for ChD (4.85%), meanwhile in 893 females, we observed 67 ChD cases (7.5%), although such difference was not significant. It is noteworthy that only the age group between 11 and 40 years had a significantly higher prevalence than the other groups, including the elder.  We observed that 9.1% of the population belong to the higher SES level, while 49% to the middle class, 36.3% to the poor, and 4.9% to extreme poverty. The weighted OR and CI resulted in no significant differences between SES and ChD, even after analyzing the four strata of SES as "dummy" variables. For its part, the SL in 1214 individuals showed a percentage of people with a null school of 31%, with incomplete primary education 50%, and very few individuals with professional or higher levels (5.6%). We found a discrete nonlinear correlation between SL and SES, exclusively in people with higher education (R 2 = 0.0216, p= 0.06). SL neither showed association with ChD. Regarding the peoples´ occupation we did not find any significant correlation with ChD, although the Homeworkers showed a slightly higher prevalence but statistically no significant.
In table 2, we observe the correlation of ChD with environmental variables. The type of settlement (rural or urban), including Guadalajara, Tepic, and Puerto Vallarta with more than 500000 inhabitants, showed no significant association. Seventy hundred eighty-two individuals have animals at home, especially pets (54.4%), although this did not show association with ChD. Concerning the quality of the dwelling, the construction characteristics with a supposed higher risk of transmission showed a slight negative relation with the prevalence of ChD, although this figure did not reach statistical significance.  Table 2: Environmental characteristics of settlements and houses of sampled people at the 58 localities and their association with ChD seroprevalence analyzed using Logistic Regression.
After that, we performed a multivariate analysis of the most influential variables transformed in a dichotomous fashion which results are depicted in Table III. For this analysis, we used the age group of 11 to 30 years as a reference. As we can appreciate, the interaction of young females living in rural areas and with a low SES, showed the most significant association with ChD.

Discussion
The ChD prevalence found by us is consistent with previous findings in the area [8,14], and it is higher than the estimated for the national seroprevalence reported by the Mexico's Secretary of Health of 0.53/ 100,000 in 2010 [15]. This fact indicates a relevant sub-register of ChD in the Mexico´s official figures.
Concerning the prevalence of ChD and demographic variables, we found a significantly higher rate in young women (aged between 11 and 30 years) living in rural settlements and with a low SES, independently from the other factors. This occurrence could depend on more intense exposure of women to infection with TC due to being at home for longer, where triatomines are colonizing with a rising frequency. Such phenomenon is similar to other observations in South America [16], and warrants further investigation since Tc infection in women, particularly of childbearing age, represents an additional challenge of public health, given the risk of congenital transmission of ChD [17].
Regarding the higher prevalence of ChD among people between 11 and 30 years old, our data contrasts sharply with former studies in South America [16], and other parts of Mexico, where authors report a median age older than 50 years [18]. We do not have a clear explanation for this higher frequency in young people; however, we can assume that prevalence of ChD in older patients is reduced because the disease produces a significant lethality after 20 years of infection [19], especially in endemic areas without access to a specific treatment, such as Mexico [20]. Therefore, we probably do not have more old patients with ChD because of mortality and not because of a protective effect of age. Out of these findings in young female population, we did not encounter significant association of ChD with SES, SL, with rural communities, poor house quality, nor with animals at home, as is generally asserted in other regions [2][3][4][5]18]. Instead, our results are consistent with recent data on ChD in the United States, where women with the highest economic level have the highest risk of infection [17]. The distribution of SES and SL in our population are similar to the distribution of economic strata amongst Mexican society according to the AMAI [12] , so we cannot attribute this finding to a SES selection bias, differing from those reports made in South America in which authors usually sampled low-income people preferably [5,16]. Our results suggests that young people, especially housewives living in recent human settlements and deforested areas, are more likely to have unnoticed contact with naturally infected triatomines present in the region, particularly Meccus longipennis, T. barberi and M. pallidipennis that are in a process of adaptation to human habitats [21]. Independently of meager houses or extreme poverty.

Conclusions
ChD in Western Mexico is a persistent public health problem both in urban and rural communities, which preferentially affects young women of low socioeconomic status: However, in older people, the prevalence of ChD is similar between individuals with any SES and SL level or house quality. Our results support the