Eco-epidemiologic prole of an area of a Chagas disease outbreak: the risk of oral transmissions in the Brazilian northeast

Chagas disease is a neglected tropical disease strongly associated with low socioeconomic status, affecting nearly 8 million people – mainly Latin Americans. The current infection risk is based on acute case reports, most of which are typically associated with oral transmissions. In the semi-arid region of Northeastern Brazil, serious outbreaks of this transmission type have surged in the last years. One of those occurred in the city of Marcelino Vieira (2016), in the state of Rio Grande do Norte. Rural residents of four municipalities surrounding Marcelino Vieira ingested sugar cane juice – which was probably ground with Trypanosoma cruzi infected insects. The structure of domiciliary unities (DUs) in the rural area of Marcelino Vieira was investigated to better understand the factors related to the outbreaks in this region – which was combined with entomological indicators. We found triatomines (mainly Triatoma brasiliensis) in 54% (36/67) of DUs and all rocky outcrops inspected (n = 7). Overall, 26% (119/458) of examined T. brasiliensis were infected by T. cruzi in articial ecotopes, with almost the same prevalence in the sylvatic environment (23%; 35/154). The local variation in T. cruzi prevalence (variating from 0%-100%) was highly correlated with the presence of some ecotopes where the insects were found; and we identied those linked to high natural triatomine infection prevalence by T. cruzi (mainly wood/tile/brick piles). Ninety-ve percent of people interviewed recognized the triatomines and knew the classic route (vector-borne) of transmission of disease. However, only 7.5% admitted knowledge that Chagas disease can also be acquired orally – which poses a risk this transmission route currently recognized. Here, we highlight the physical proximity between humans and infected vector populations as an additional risk factor to oral/vector contaminations, providing recommendations to avoid further outbreaks.


Introduction
The protozoan parasite, Trypanosoma cruzi, is the etiologic agent of Chagas disease -one of the most relevant neglected tropical diseases in Latin America (WHO, 2020). Classic transmission occurs via excreta from blood-sucking bugs that are eliminated during or shortly after the blood meal of infected insects.
These blood-sucking bugs belong to the Triatominae subfamily (Hemiptera: Reduviidae; Lent and Wygodzinsky, 1979). Triatoma infestans, the primary vector, was likely introduced from the Bolivian Andes and adapted to households across Brazil -probably during a period known as "the sugar age" (1530-1700), during the exploration of sugar cane, coffee and ores in Colonial Brazil (Gurgel et. al. 2009). The Southern Cone countries (Argentina, Brazil, Chile, Uruguay, Paraguay, Bolivia and Peru) joined efforts in 1991 to combat transmission by T. infestans and to control the transmission via blood/organ donations (Schmunis and Yadon, 2010). As a result, Brazil became practically free of T. infestans since 1996; and has eliminated transfusional Chagas disease transmissions nearly completely (Dias, 2007). However, native vectors, such as T. brasiliensis and T. pseudomaculata, currently challenge control measures by invading and colonizing human domiciles in endemic regions in northeastern Brazil (Dias et al., 2000;Lima et al., 2015). Additionally, outbreaks by oral contamination have been reported in several Brazilian regions (Dias et al., 2016). The oral route is recognized currently as the main source of infection, especially in the Brazilian Amazon, and most likely the result of the consumption of fresh açaí berries and juice in this region (Dias et al., 2001).
Chagas disease outbreaks in the Amazon region have been intensively investigated both from epidemiological, ecological and molecular perspectives. Such work has led to the genotyping of the parasite from vectors, T. cruzi reservoirs and humans (e.g. Coura  The results of this study suggest that T. tibiamaculata was the responsible vector (Steindel et al., 2008). In the municipality of Guarapari, in the state of Espírito Santo (southeastern Brazil), an oral case of Chagas disease transmission resulted in the death of a two-yearold child (Dario et al., 2016). From this incident, the sylvatic (rodents) and domestic (dogs) reservoirs of T. cruzi were suggested; and the parasite was genotyped (DTUs I, II, III and IV) in the human fatal case, which was the same genotypic combination harbored by the vector (T. vitticeps) (Dario et al., 2016).
Currently, vector control measures in the Brazilian Northeast depend on expensive governmental initiatives by states and municipalities. Governmental initiatives are needed to support regular, although costly, intervention to monitor and control T. brasiliensis and T. pseudomaculata infestation if low infestation rates are desired. As an additional complication, the Brazilian Northeast is a semiarid region that is plagued by periodic drought and lack of economic resources. These factors exacerbate the already taxed government resources, and result in discontinuity of vector control operations (Dias et al., including the potential for oral infections. To determine additional factors possibly linked to the outbreak, the distribution of infected/non-infected insect populations and their foci -in both sylvatic and peridomestic ecotopes -were mapped.

Materials And Methods
2.1 Ethics statement. This research was approved by the UNICAMP Research Ethics Committee (protocol No. 2,631,532). The collection and transportation of triatomines were conducted with the assistance of technicians from the municipal and state health departments and had the Sisbio IBAMA License N o 58373-1 approval. We obtained permission from homeowners/residents to collect insects from all dwellings and properties, and all interviewed residents signed (or printed digitally) a Free and Informed Consent Form (FICF). The SISGEN register is A5C8D0D.
2.2 Study area and environments. We conducted the study in the municipality of Marcelino Vieira (6° 17' 38" S, 38° 10' 01"W), located within the Caatinga biome in Northeastern Brazil. The region is characterized by seasonally dry forests with a mosaic of thorny shrub vegetation (IBGE, 2017). We chose this site because (i) it had the most recorded infections of the outbreak, and (ii) it is the location of the sugar cane mill -the potential source of the outbreak. Inspections for triatomine infestations were conducted in the following environments: the sylvatic (rocky outcrops), the peri-domestic and the intra-domestic.
The term domiciliary unit (DU) here refers to the combination of peri-domestic and intra-domestic environments.
2.3 Surveys with rural property residents. We conducted a cross-sectional study through interviews with residents of 67 DUs four rural localities of Marcelino Vieira by using a questionnaire. These DUs included patients affected by the Chagas disease outbreak. One resident (over the age of 18) from each dwelling was interviewed to obtain socioeconomic and educational information. Speci cally, we collected data on age, gender and years of schooling of residents. We grouped residents into the following age groups: 18-30, 31-59 and over 60 years. To establish the socioeconomic characteristics, we followed the Brazilian Association of Research Companies (ABEP), which is based on a proprietorship score and level of education. We classi ed the total income per household based on Kamakura and Mazzon (2016), but it was adapted to the reality of the local residents from the rural and poor site studied, as follows: monthly family  To test vector recognition, we presented each respondent with a selection of immature and adult triatomines and non-hematophagous insects (e.g., beetles).

Characteristics of dwellings construction.
We evaluated the construction of each dwelling via home inspection and classi ed it by its probability of triatomine colonization. For example, whether walls were coated and whether the roof had a ceiling to prevent the insect colonization. The survey conducted here was adapted from a survey produced by FUNASA (Brazilian Health Foundation) to guide house improvement programs (FUNASA, 2017).

Insect captures and entomological indicators.
We searched for triatomines inside homes (intradomestic), which includes all areas closed by doors where humans live or work (Almeida et al., 2008). In these environments, we examined cracks in the walls, the back of paintings, mattresses, cabinets, curtains, and under/behind furniture. We also searched peridomestic environments, represented by spaces surrounding homes. In these environments, several man-made structures are found -mainly shelters for domesticated animals (e.g., chicken, goats, cows, and dogs). Besides livestock shelters, it is common to nd woodpiles and brick/tile piles (i.e., arti cial ecotopes) to store material for rudimentary ovens and to build houses, walls, roads, fences and other purposes. We took note on each ecotope observed and investigated.
We also examined the primary and native sylvatic ecotopes of T. brasiliensis (e.g., rocky outcrops; Lent and Wygodzinsky, 1979) as well as alternative natural ecotopes (e.g., xique-xique cacti, lower tree trunks, downed trees, and animal dwellings; (Valença-Barbosa et al., 2014). We identi ed these locations using Google Earth Pro software (Wuthrich, 2006) or by consulting the residents. We divided sampling into four groups of localities (A, B, C and D) -distributed according to their geographic proximity. Sylvatic points that were not around DUs were also considered, maned "O" (outside localities; Figure 1). We performed searches beginning at sunset (~17:00 local time) and continued collection for approximately ve hours. All insects were collected using cap lamps, gloves, and tweezers. 2.8 Trypanosoma cruzi natural infection. We developed our procedure using traditional optical microscopy techniques (OM). We obtained triatomine feces using abdominal compression and deposited the sample on a slide, adding 5µL of saline solution. We then covered the slide with a 22 x 22 mm glass coverslip. We examined the intestinal content using a binocular optical microscope with a 400x magni cation by screening all elds for T. cruzi. We conducted the molecular diagnosis simultaneously using Polymerase Chain Reaction (PCR), for comparative purposes, to assure that the trypanosomatids observed were T. cruzi. We macerated the gut of a portion of collected insects using a sterile crusher and liquid nitrogen. We then used The DNeasy Blood & Tissue Kit (Qiagen) to isolate the DNA, following the manufacturer's protocol. We used PCRs the primers 121 and 122 (T. cruzi kDNA), as previously described by Sturm et al. (1989) and Wincker et al. (1994), with termocycles described by Moreira et al. (2017). For all insects that tested positive for T. cruzi, we ampli ed the 330 bp fragment. If positive for T. cruzi, an amplicon could be seen in 2% (w/v) agarose gels by electrophoresis at 80 V for 40 min. We used samples with water as controls for comparison with the DNA Dm28 strain from T. cruzi culture (TcI). An additional PCR was conducted to verify intestinal DNA did not have inhibitors for ampli cation; therefore, for the same samples, we ampli ed the CytB gene as the target. For this purpose, we ampli ed the insect Cyt B gene as the target, by using the cybTprR /CYTBR (Oliveira et al., 2017) and 7432F primers (Monteiro et al., 2003), with PCR conditions previously described.

Ecotypic indicators. The Ecotypic Prevalence in
2.9 Statistics. We used chi-square analysis to evaluate the signi cance among entomological indicators obtained among localities and McNemar's chi-square to compare traditional and molecular techniques for T. cruzi detection. We evaluated statistical signi cance using 95% con dence intervals and an alpha of 0.05. All analyses were conducted using the R (R Development Core Team, version 3.6.3; https://www.r-project.org/).

2.10
Relationships among Trypanosoma cruzi infection, habitat and localities for Triatoma brasiliensis. To explore the interaction among habitat sources, T. cruzi prevalence and localities, we used the software Cytoscape 3.7.2 (Shannon et al., 2003) to build a network. For this purpose, we used only insects that tested for the T. cruzi natural infection via MO.

Results
3.1 Surveys with rural property residents. We interviewed 67 people -composed of one person per UD inspected. Respondents were varied among age groups: (i) 17.9% were 18-30 years old, (ii) 43.3% were 31-59 years old and (iii) 38.8% were 60 years old or older. Most respondents were women (76.1%). The majority of people also had less than ve years of schooling or were illiterate (Table 1). For the ones who have a formal monthly salary, the main source of income derived from retirement related to age or disabilities (55.2%), governmental social programs (6.0%), from livestock and subsistence farming activities (31.3%) and other activities 7.5%. According to socioeconomic parameters, families should receive a minimum monthly of USD $ 197.91 (see: https://www.gov.br/; dollar quotation in September 16, 2020); but 25% lived with on less than USD $40.5 Therefore, family income is mostly dependent on self-subsistence agricultural activities (Table 2). Regarding the knowledge of the residents on Chagas disease, 95% of those interviewed recognized the insect vector, knew the ecotope it occupies in DUs, and most of the respondents (92.5%) understood that the triatomine transmits a disease. Only 7.5% reported that they knew the disease could be transmitted by contaminated food/drink (oral route). Thirty percent of the interviewees said they believe to have been bitten by the vector in the last 3 years. (Table 3). Most of them supposed to have been bitten by the nding of blood-engorged insects around where they were sleeping (e.g. under beds, behind furniture, curtains, etc.). Do you think you have had contact with "barbeiros" or have been bitten by these insects in the last 3 years? 20/67 30 A respondent may have provided more than one answer. *People who answered they knew the disease is transmitted only via the insect´s bite The main sources of information on Chagas disease came from within the community (i.e., neighbors) according to 49.3% of respondents (Table 4).  . 2003) and was more abundant, in this study, we will only report on ndings associated with this species.
Ecotopes. We measured the prevalence for the nding of key-ecotopes (Ecotypic Prevalence in DUs: EPD:) and the proportion of those with triatomine infestations (Prevalence of Ecotypic Colonization, PEC). Overall, henhouses were the most common ecotope found in DUs (EPD=85%). However, we found greater insect colonization in woodpiles (PEC=47.1%). Both EPD and PEC varied among localities (Table 5). For both indicators, the distribution was signi cant among localities at p <0.05 (see details in topic 3.4.1) 2%. These indicators varied among localities; and the most remarkable variation was in the NI, ranging from 0% (A) to 100% (C). Localities B and C were slightly closer to rocky outcrops and exhibited higher DI, DTD and Domiciliary Colonization Index (DCI). A high variation was also found for DTD, ranging from 3.1 (A) to 16.9 (C) ( Table 6). For all indicators, the distribution was signi cant among localities at p <0.05 (see details in topic 3.4.1)

Entomological indicators for sylvatic environments.
Overall, we collected 1,096 insects in sylvatic environments, varying from 29 (C2) to 287 (D). We had only a few specimens for which to test the natural index of T. cruzi infection in the sylvatic environments (NIS) because most of the insects either were too emaciated to be analyzed or died shortly after collection. As happened for indicators in anthropogenic environments, the natural infection also ranged among sylvatic environments, even within the same locality. For example, in the locality A, where no peridomestic insects were infected; however, we found 100% (10/10) infection in bugs from the sylvatic environment. In the sylvatic habitat, overall triatomine density (TDS) was 9.7. TDS ranged from 1.8 (C2) to 17.9 (D) ( 3.5 Comparison of Trypanosoma cruzi detection on insects via traditional and molecular methods. We randomly selected 159 samples of T. brasiliensis for PCR analysis to identify infection by T. cruzi. We based the criterion on ecotypic/geographical balancing. The number of samples varied across sampling spots within localities, such that more samples were included from areas with greater epidemiological. For example, we analyzed 30 insects from the site of the sugar cane mill (and probable outbreak site; population "vii" in locality B) and 34 insects from a site with a numerous infestation (n= 191 insects collected), compared to other spots with lower number of samples caught (n< 70). Insects from populations "vii" and "ix" had the highest T. cruzi prevalence, considering their sample size, using both Optical Microscopy (OM) (93% and 100% respectively) and PCR (100% for both populations) methods (Table 8). We found no statistical difference in prevalence identi ed by the OM and PCR methods (χ 2 = 3.03, df = 1, p = 0.081). 3.6 Relationships among Trypanosoma cruzi infection, habitat and localities for Triatoma brasiliensis. We analyzed a total of 612 specimens using the OM method. From this analysis, we demonstrate the relationship among the T. cruzi prevalence, ecotopes, and localities. At the locality, A, 21.8% (60/275) of the analyzed sample was collected in henhouses; the remainder of the sample originated from woodpiles in the peridomestic environment (all of which were negative for T. cruzi natural infection) and from rocky outcrops (n=10) -which were all positive for T. cruzi. Except for henhouses, the remaining peridomestic habitats (woodpiles, tile piles and woodpiles combined with tile piles) exhibited high proportion of infected insects (>45%). The presence of tiles seems to be more likely to shelter T. cruzi infected bugs, as higher T. cruzi prevalence is observed for tile alone and for the combination of woodpiles + tile piles. Infected insects were identi ed in all localities that had rocky outcrops, with an overall 25.5% ( 35/154) (Figure 2).

Discussion
Although vector-borne Chagas disease transmission has declined dramatically in Brazil since 1980, acute cases of the disease continue to be reported . In this study, the permanent income of surveyed residents was di cult to assess, since the income of rural workers is not permanent due to the productive cycles of crops and animals.
Awareness of transmission risk and mode may also affect infection rates among residents. Most of the respondents in our study reported that they understood the classic vector-borne transmission; however, only a small percentage of these respondents (7%) were aware of the risk of transmission through oral contamination. According to Colosio et al. (2007), in Paiçandu-PR, the greatest knowledge gap in the population was a contributing factor to disease transmission.
Previous research suggests that dwelling structures may create habitat for, and increase colonization risk of, triatomine infestations because walls without coating can work as shelters for triatomines in their cracks, and the roofs without ceilings present a direct pathways between the house and the external area, facilitating the circulation of the vector. We found that 12% of the houses are without coating or ceilings, which is slightly higher than houses in both urban The localities with the highest values of wild triatomine density were also those with the highest dwelling infestation, suggesting that triatomine density may act positively on infestations. Regarding the distances between rocky outcrops and DUs, almost all localities had sylvatic environments within the estimated range of ight capacity (Cecere et al., 2004) for triatomines (~500m). The exception was for the locality D, but the number of houses survived (n = 5) was insu cient for robust inferences on the role of sylvatic foci as source of domiciliary infestation.
Chicken coops were present in 85.1% of the houses inspected, but of this, only 31.6% was infested. Tile, brick and woodpiles were less prevalent in in DUs (25.4-45%) but 43-47% of them were infested. Additionally, insect populations collected in these last mentioned ecotopes exhibited much higher T. cruzi prevalence, which was evidenced in the network. It is common to nd piles of building materials and wood in proximity to structures and dwellings. These piles are readily used as arti cial ecotopes by triatomines, as also observed by (Coutinho et al., 2012) in Ceará. According to Lilioso et al. (2020), these ecotopes can shelter potential T. cruzi reservoirs.
There are many factors that in uence triatomine infestation in DUs (e.g., proximity of houses to rocky outcrops, debris piles near houses, and proximity of domesticated animals to living spaces). Although most people cannot relocate their houses further from the natural habitat for triatomines, educating residents about other factors in their control may reduce local infestation (Dias et al., 2016). That is, understanding the relationship between infestation and ecotopes may reduce transmission disease risk. We found that, with the exception of henhouses, the prevalence of infected insects in peridomestic environments was even larger than that for the sylvatic environment. Thus, educational programs could be used to inform residents about the risks associated with creating arti cial ecotopes in the peridomestic environment.
We found no difference between the technique used to detect T. cruzi-infected triatomines. There as only an 8% difference between PCR and OM. , where the prevalence may be more variable (2.9-20.3%). These high T. cruzi circulation in Rio Grande do Norte State was also evidenced by Vicente and Camara (2019), who identi ed 40% (16/40) of analyzed dogs positive for T. cruzi. These studies underscore the risk of T. cruzi transmission to humans and the need to monitor T. brasiliensis domiciliary infestations to prevent outbreaks -particularly in the Rio do Norte state.
The outbreak extended to three municipalities beyond Marcelino Vieira: Tenente Ananias, Alexandria and Pilões (Vargas et al., 2018). These municipalities are located in the southwestern region of the state and are within a distance of 17 to 30 km from each other. We believe that the extent of the outbreak was underestimated because it was based on symptomatic patients that sought health assistance. Futhermore, the number of respondents that had contact with triatomines raises concern and should be a focus of future monitoring efforts if health o cials seek to interupt transmission. The results of our study highlight the need for a human serological survey in the affected area, where peridomestic triatomine infestation by T. cruzi infected insects is high. According to WHO (2020), widespread diagnosis is required to treat patients infected by T. cruzi in the speci c phase of the disease (acute and chronic), and in a timely fasion.
Concluding remarks. Thirteen percent of people reported they may have been bitten by triatomines in the last three years. Twelve percent of dwellings were found to be suitable for triatomine colonization (e.g. those living in a dwelling without coating or covered roof). Educational programs may change the way residents deal with spaces surrounding homes to avoid creating sites that attract T. cruzi infected insects. The proximity of infected bugs and humans is worrisome for both vector-borne and oral transmissions. We demonstrated that the population sampled in our study has all the characteristics of those affected by Chagas disease: low income, low level of education, and a willful disregard for the routes of Chagas disease transmission (speci cally oral transmission). We reported that information about the disease usually comes from word of mouth in the community and still covers only the classic transmission route (via insect bites). All of these factors together present a situation that leads to increased risk of infections or new outbreaks because this scenario is probably similar for other points non investigated. Oral route transmission a will be a challenge to public health programs working to combat Chagas disease; however, such programs that speci cally target oral transmission and include community education and participation must be funded and administered in Northeastern Brazil. The funders had no role in study design, data collection, analysis, decision to publish, or preparation of the manuscript.