Triatomine infestation and vector-borne transmission of Chagas disease in northwest and central Paraná , Brazil

Triatomine infestation, prevalence of T. cruzi antibodies in humans and domestic animals, and variables potentially associated with the presence of triatomines in a rural domiciliary unit (DU) were investigated in nine municipalities and one district of Paraná, Brazil, from June 1996 to February 2000. DUs were defined as all houses and annexes on a given piece of property. Blood samples from human volunteers, dogs, and cats were submitted to indirect immunofluorescence. An epidemiological form was completed for each DU. A logistic model was adopted in order to identify associations. Seven out of nine municipalities were positive for triatomines. T. infestans was not captured, but T. sordida, P. megistus, and R. neglectus were. Different variables were considered decisive for the presence of triatomines across the municipalities: proximity to residual wooded areas, i.e. either scrub forest (capoeira) or islands of residual forest (slightly dense vegetation), longer time of residence, existence of inhabited houses, and past DU infestation. In order to ensure proper continuation of the recommended Chagas disease control program, entomo-epidemiological surveillance measures need to be maintained. Triatoma; Epidemiologic Factors; Chagas Disease Introduction Over the last thirty years, Chagas disease transmission has declined significantly both in Brazil and in a large part of Latin America as a whole, as a result of vector control through the use of insecticides and improvements in rural housing, living standards, and education 1. In Brazil, rural exodus in the 1970s and ’80s reduced the use of manual labor in rural areas, thereby increasing the number of inhabited houses, which provide ideal shelter for Triatoma infestans Klug, 1834 2. In Brazil, the elimination of triatomines from domestic environments does not avoid the persistence of residual foci, particularly in the peridomicile 3, due to the invasion and colonization of artificial ecotopes by triatomines whose natural habitat is the sylvatic ecotope. Such triatomes include Panstrongylus megistus (Burmeister, 1835), Triatoma sordida (Stal, 1859), Triatoma braziliensis (Neiva, 1911), Rhodnius prolixus (Stal, 1859), and others 4. Camargo et al. 5 reported that the state of Paraná, Brazil, had an elevated number of seropositive individuals for Chagas disease, placing it fourth among all Brazilian States in a serological survey to determine Chagas infection. Until 1995, current data were scarce in the literature in relation to secondary triatomine species in Paraná 6,7,8,9. Toledo et al. 10 found that 98.7% of the triatomines collected in rural areas of Falavigna-Guilherme AL et al. 1192 Cad. Saúde Pública, Rio de Janeiro, 20(5):1191-1200, set-out, 2004 municipalities in northwest Paraná belonged to the Tr. sordida species. However, Silveira & Vinhaes 11 did not report any triatomines in this State during an investigation carried out by the National Health Foundation (FUNASA) from 1995 to 1997. Furthermore, Falavigna-Guilherme et al. 12 analyzed the majority of municipalities from the present study and found that uninhabited houses were the most frequently infested ecotopes, followed by chicken coops and inhabited houses. The lack of current information on Trypanosoma cruzi infection and triatomine behavior in the domiciliary and peridomiciliary environments motivated the present investigation on triatomine infestation, prevalence of T. cruzi antibodies in humans and domestic animals, and variables possibly associated with the frequency of triatomines in such environments in Paraná State. Material and methods


Introduction
Over the last thirty years, Chagas disease transmission has declined significantly both in Brazil and in a large part of Latin America as a whole, as a result of vector control through the use of insecticides and improvements in rural housing, living standards, and education 1 .In Brazil, rural exodus in the 1970s and '80s reduced the use of manual labor in rural areas, thereby increasing the number of inhabited houses, which provide ideal shelter for Triatoma infestans Klug, 1834 2 .
Camargo et al. 5 reported that the state of Paraná, Brazil, had an elevated number of seropositive individuals for Chagas disease, placing it fourth among all Brazilian States in a serological survey to determine Chagas infection.Until 1995, current data were scarce in the literature in relation to secondary triatomine species in Paraná 6,7,8,9 .Toledo et al. 10 found that 98.7% of the triatomines collected in rural areas of municipalities in northwest Paraná belonged to the Tr.sordida species.However, Silveira & Vinhaes 11 did not report any triatomines in this State during an investigation carried out by the National Health Foundation (FUNASA) from 1995 to 1997.Furthermore, Falavigna-Guilherme et al. 12 analyzed the majority of municipalities from the present study and found that uninhabited houses were the most frequently infested ecotopes, followed by chicken coops and inhabited houses.
The lack of current information on Trypanosoma cruzi infection and triatomine behavior in the domiciliary and peridomiciliary environments motivated the present investigation on triatomine infestation, prevalence of T. cruzi antibodies in humans and domestic animals, and variables possibly associated with the frequency of triatomines in such environments in Paraná State.

Study area
Fieldwork was carried out in the rural areas of nine municipalities in Paraná that are located mainly in the northwestern region of the State.The northwest Paraná meso-region is situated between 52º20' and 52º50' longitude West and 23º20' and 23º60' latitude South.The region is part of the geomorphological unit called the Paraná Basin Central Plateau 13 , which is part of the Third Paraná or Sandstone-Basaltic Plateau.The terrain consists of low, flat-topped plateaus with convex slopes that reach altitudes of 250 to 600m.The climate is classified as warm, super-humid tropical 14 .Average annual rainfall is 1,700mm.In the dry winter season, average temperatures are below 18ºC, while the average yearly low is 16ºC and the average annual highs range from 26 to 28ºC 15 .The overall annual temperature is 21.9ºC.
Nearly 95% of the original vegetation in the northwestern region of the State, consisting of latifoliate subtropical forest, was destroyed to make way for cropland 16 .The region's current economy is dominated by production from small and medium-sized farms.The farming industry consists mainly of feedlot production and crops including soybeans, corn, sugarcane, manioc, wheat, and grapes 15 .The rural exodus from northwest Paraná has been intense since the 1970s 16 .Many dwellings from earlier decades that had been linked to coffee growing were torn down, while most of the remaining structures were abandoned.Today, approxi-mately 90% of the existing houses are built of wood with ceramic tile roofs and have electricity.
In the north-central Paraná meso-region (or second plateau), only the municipality of Cândido de Abreu was studied.Despite vast wastelands, the region has small areas of mixed vegetation including Brazilian pine and broadleafed tropical woodlands 13 .The municipality of Cândido de Abreu has an altitude of 600m and is situated at 51º19' longitude West and 24º36' latitude South.The climate is mesothermal, humid, and subtropical with hot summers, concentrated rainfall, average temperatures over 22ºC in the hottest month and 26 to 27ºC average annual highs 14 .Freezing temperatures in the winter are infrequent.Minimum temperatures are around 12.7ºC, with 14 and 15ºC as the average annual lows.There is no well-defined dry season 15 .Economic production is centered on range-fed livestock production.However, there is also crop cultivation, mainly beans, tobacco, corn, potatoes, and soybeans 13 .The municipality has extensive poverty, and mud-and-wattle houses are not uncommon in the rural area 13 .Its territory is approximately equivalent to the total of the other municipalities studied here (Figure 1).
Due to intense deforestation in the State of Paraná, what remains is a predominance of primitive, seasonal, latifoliate forest that appears in the considerably modified form of thinly-spread forest and scrub, characterized by open spaces with only sparse vegetation, punctuated by small trees and shrubs/bushes that are highly exposed to solar radiation.This description agrees with that of Forattini 17 , who portrays Paraná as an originally umbrageous area, currently reduced to mainly open spaces due to anthropic action and presenting an ugly landscape dotted with islands of residual vegetation.

Municipalities studied
Figure 1 shows the study area of eight municipalities and one district in northwest Paraná, and one municipality located in the central part of the State.
The selection of municipalities was based on work by FUNASA, guided by frequent findings of triatomines by inhabitants of these municipalities and by entomo-epidemiological data from each municipality obtained in the 1980s, classifying them as levels I, II, or III in terms of priority.According to FUNASA technical standards 18 , priority I municipalities are those where it is still possible to find Tr.infestans (the main Chagas disease vector) in the domiciliary or peridomiciliary environments.In these areas, research is conducted annually to identify triatomines.Priority II municipalities are those where FUNASA has verified the colonization of an inhabited domicile within the last three years, or where reports have been made of the presence of secondary species such as Tr.sordida and P. megistus.This classification requires that farms within the municipalities be surveyed every two years for triatomines.Priority III municipalities are those where FUNASA has identified colonization by Tr. sordida and P. megistus in peridomiciliary annexes in the previous five years, and require triatomine surveys every three years.
The municipalities of Floresta, Mandaguari, Ourizona, and Maringá, including the rural area of the Floriano district, are classified by FUNASA-Paraná as "priority II".The munici-palities of Paiçandu, Dr. Camargo, and Marialva are considered "priority III", while Cândido de Abreu is "priority I".The Municipality of Porto Rico has not been classified by FUNASA.
In these municipalities, work by FUNASA was conducted four days a week.To obtain the present data, the team from the State University of Maringá (UEM) accompanied FUNASA one day a week.This procedure meant that data collection had to occur at homogeneously distributed locations in each municipality at different points.

Procedures
The study was conducted according to the FU-NASA activity timeline for entomological surveillance or based on reports of infestation submitted to FUNASA.Triatomines and epi-

Entomological inspection
The search for triatomines in houses and annexes was carried out by a team of six workers from FUNASA and three from UEM.A "domiciliary unit" (DU) was defined as all the houses plus annexes (chicken coops, pigsties, barns, wood sheds, stables, etc.) at each of the rural properties visited (cabins, cottages, farms).An uninhabited DU was defined as a DU with no residents and was considered as a whole, without distinction between intradomiciliary and peridomiciliary environments.
Triatomines were captured manually from external and internal walls, roofs, furniture, mattresses, and other objects within the domiciles.The capture methodology could not be classified in "man/hours" or "dislodging insecticide employed" since the entire DU was searched, regardless of the time spent on it.The finding of triatomines signaled the end of the search in a given DU.Similar procedures were adopted for inhabited and uninhabited DUs and were considered positive if nymphs and/or adults were found.Infested DUs were treated by FUNASA with the insecticide cypermethrin (125mg active ingredient/m 2 ).
At least 10.0% of existing DUs were inspected in municipalities of up to 900 DUs.For municipalities with over 900 DUs, a minimum of 1.0% was inspected.

Prevalence of T. cruzi antibodies in humans
Blood samples were collected from volunteers who gave informed consent.The samples collected for this work were calculated using the computer program Epi Info version 6.0 considering the rural population of each municipality, a prevalence of 4% 5 , and 95% confidence levels.Samples were collected at all sites.They were deposited on filter paper after finger sticks with sterile, disposable lancets, dried at room temperature, and stored at -20ºC until use.All blood samples were submitted to indirect immunofluorescence (IIF) according to Camargo et al. 5 19 .

Prevalence of T. cruzi antibodies in cats and dogs
Priority was given to animals that most frequently entered wooded areas near domiciles.Blood from cats and dogs was collected onto filter paper from a small ear incision.Samples were processed by IIF according to Camargo et al. 5 using specific anti-dog and anti-cat Sigma conjugates and Biolab antigens.

Epidemiological data collection instrument
The form utilized was based on the model used by Diotaiuti et al. 3 .During site visits, one form was completed for each DU, whether inhabited or not.Data included type of dwelling, number of residents, length of residence, presence of annexes and distance from the domicile, presence of animals, distance to closest DU, distance of DU from residual wooded areas in general ("scrub forest" or "islands of residual forest"), and presence of triatomines.

Statistical analysis
Data from the epidemiological forms, entomological inspection, and serological results were entered into a database using the Statistica Program version 8.0.Since it was binary, a logistic model was adopted to describe the response 20 .The clustered K-averages model was used to categorize the quantitative explanatory variables.Statistical analysis of the information collected from each municipality consisted of describing the characteristics of the DUs, identifying associations, and obtaining a model that would explain the presence of triatomines.

Research Ethics Committee approval
The study protocol was approved by the UEM Institutional Ethics Committee.All study participants provided informed consent.

Results
Different localities were investigated in each municipality (841/143) (Table 1), corresponding to 17.0% of existing sites.Triatomines were found in seven of the nine municipalities analyzed and in the Floriano district (Table 1).Of the captured triatomines, 92.4 % were found in houses and annexes built of wood.
As shown in Table 1, 522 samples of the Tr.sordida, P. megistus, and R. neglectus species were collected, of which 370 were nymphs (70.9%) and 152 (29.1%) adults.The majority of the specimens were Tr.sordida, which represented 86.6% (329 nymphs and 123 adults) of all the triatomines collected, mainly in the peridomicile.P. megistus (41 nymphs and 27 adults) was the second most common species (13.0%).Only two specimens of adult R. neglectus were identified.
Triatomines captured in the intradomicile of inhabited DUs were mainly adults (nine Tr. sordida, five P. megistus, and two R. neglectus).Five nymphs captured in the intradomicile in Maringá and Floriano were Tr.sordida.
Serological testing was performed on 4.2% of the rural population within the study's catchment area.If only the northwest region of the State is considered, i.e. excluding the Municipality of Cândido de Abreu, serology was performed on 6.0% of the 26,215 inhabitants in the rural area Samples were collected from DU residents present at the time of the visit, and no one refused to cooperate with the procedure.Humans and animals whose blood samples were taken belonged to DUs from localities which were homogeneously distributed throughout the municipality.Table 3 includes the results for prevalence of T. cruzi antibodies, with testing performed on 1,762 (61.8%) humans, 21.0% of whom were less than 15 years old, all negative.Ten subjects 15 years or older were seropositive for T. cruzi, and they had the results confirmed in the test conducted in serum.These patients were referred to the Health Service for diagnostic confirmation and to the ACHEI Program for psychosocial support 19 .Nine of the ten seropositive individuals were originally from the State of São Paulo, and one was from the State of Minas Gerais.All reported that during their childhood in those States, they had lived in houses infested with barbeiros (the common name for the vector).These subjects reported that while residing in Paraná they did not recall having seen triatomines in their homes.As for prevalence of T. cruzi antibodies in domestic animals, none tested positive (Table 3).
Of a total of 1,227 dogs and 979 cats on inhabited rural properties visited during the course of the study, serological testing was performed on 122 (9.9%) and 49 (5.0%), respectively (Table 3).None of the domestic animals was seropositive for T. cruzi.
According to the logistic model, in the Municipality of Paiçandu there was a significant association (P < 0.05) between presence of triatomines and longer time of residence (16.6 ± 14.3 years) and distance between the DU and residual wooded areas or "scrub forest" (441.2 ± 434.5m).However, only distance between DU and residual wooded areas entailed a significant risk (1.27) of triatomine infestation (the closer the dwelling to woods, the higher the risk).For the other municipalities, the logistical model was not sensitive enough to calculate the risk of infestation, due to the small number of triatomines captured.In the municipalities of Mandaguari, Maringá, and Cândido de Abreu, the existence of inhabited houses was the determining factor for triatomine infestation in the logistic model, reinforced by past infestation of the DU.In the municipalities of Floresta, Marialva, and Ourizona and the Floriano district, triatomine infestation at the time of the study was determined by past infestation of the DU.On the other hand, triatomine infestation in the Dr. Camargo municipality was associated with the variables past DU infestation, presence of an uninhabited DU, and proximity of a "vacant dwelling" annex to the domicile.The number of peridomiciliary annexes (3.6 ± 1.5) and proximity of DUs to residual "scrub forest" (275.2 ± 336.6m) also helped explain the presence of triatomines in that municipality.
DU infestation with triatomines in the Municipality of Porto Rico was not associated with any of the study variables in the logistic model.The species Tr. infestans was not found during the course of the current study.Three species of triatomines -Tr.sordida, P. megistus, and R. neglectus -were infesting DUs in seven of the nine municipalities investigated, justifying their classification as priority II and III at that time by FUNASA.
Tr. sordida was the most frequent species in the municipalities, particularly in the peridomiciliary setting in Paiçandu, where 36.7% of the inhabited DUs were infested.It is worth noting that the higher the degree of DU infestation, the greater the possibility of intradomiciliary infestation.However, this pattern is not a rule.In Mambaí, Goiás State, this species was found in more than 10.0% of DUs, in both the peridomiciliary and intradomiciliary environments 22 .In Porteirinha, Minas Gerais, 34.9% of DUs were infested with Tr. sordida, mainly in the peridomiciliary habitat 3 .The State of São Paulo is at present directing its control towards the secondary species Tr. sordida, which is also the prevalent species found in the peridomiciliary environment 23 .According to Forattini et al 24 , Chagas disease occurs endemically in Brazil only where there is the possibility of vector domiciliation, and the role of Tr. sordida in T. cruzi transmission is negligible when there are low levels of DU infestation and when colonization is lacking.
A lower proportion of the P. megistus species was collected only in the municipalities of Maringá, Mandaguari, Marialva, and Porto Rico, mainly in the intradomicile of inhabited DUs.In the first three municipalities, this work was temporarily suspended due to logistic problems related to FUNASA policies and priorities.
No triatomines were captured in Cândido de Abreu during the course of the present study.Nevertheless, the municipality has been included in the Triatoma infestans Eradiction Program -which classifies municipalities as free of Tr. infestans when they present totally negative DUs for a period of three consecutive years 21 -since this species may still be cap- tured in this municipality.On visits made after the collection period of the present study, FU-NASA captured P. megistus from DUs in Cândido de Abreu and Paiçandu.According to the literature 4,11 this species has a higher T. cruzi vector potential than Tr.sordida.
Based on observation, in the various municipalities Tr. sordida was related to residual forest or "scrub forest" and P. megistus to the "islands of residual forest".These data are consistent with the literature 17,25 .
A large portion (49.1%) of P. megistus specimens captured were which did not present colonization, although the research was conducted outside this species' dispersion period (Table 1) 24 .For Tr. sordida, adults accounted for only 27.2% of the specimens captured (mainly in the peridomicile).This provides evidence of colonization by this species.
Although intradomiciliary infestation in the various municipalities occurred by Tr. sordida nymphs or P. megistus adults and no positive individuals under 15 years of age were diagnosed, abandonment of the Control Program by normally responsible individuals is of great concern.This is because Tr. sordida colonies were detected in the peridomicile, P. megistus displays great T. cruzi transmission potential, and according to Falavigna-Guilherme et al. 12 the T. cruzi infection rate for Tr.sordida and P. megistus captured in these same municipalities was 13.4% and 13.5%, respectively, although human blood was only detected rarely as a meal source for these triatomine species.In addition, the chance occurrence of three isolated cases of vector-borne disease in children less than twelve years of age in Paraná in a two-year period must be considered, since they were diagnosed under special conditions 26 .
It is important to note that many DU owners reside in urban settings and frequently transport objects (including timber) from the rural to the urban area and vice versa.In addition, most rural residents do not express concern over the unkempt state of the peridomiciliary environment and many do not appreciate the importance of the presence of triatomines and the problems related to Chagas disease.In addition, some DUs are temporarily uninhabited due to the type of crops grown on the farm property but may receive workers who stay overnight during the harvest season, in proximity with rodents, marsupials, and triatomines originating from nearby "scrub forests".
The factors contributing to the presence of triatomines in the DUs were the same in all nine municipalities.Triatomine presence in DUs in the Municipality of Paiçandu, according to the logistic model, showed that "proximity of the DU" to "scrub forest" was significant.Furthermore, the longer the "length of residence" in the Municipality of Paiçandu, the higher the possibility that the DU was infested with triatomines.This is in accordance with the characteristics of Paiçandu residents who had lived there the longest and had the habit of storing debris in the peridomicile.
According to the logistic model, the presence of triatomines in DUs in Mandaguari, Maringá, and Cândido de Abreu was closely associated with the "interior of inhabited domiciles".This trend agrees with P. megistus habits and environmental features in Mandaguari and Maringá, where intradomiciliary P. megistus-positive DUs were located near "islands of residual forest".In Cândido de Abreu, the failure to find P. megistus may have been related to the low percentage of DUs visited, but it might also have been related to the intensification of FUNASA's work in recent years, contributing to triatomine control in this municipality.
Triatomines were found primarily in the peridomicile in the municipalities of Floresta and Marialva and the Floriano district.Most specimens captured belonged to Tr. sordida.In these municipalities, according to the logistic model, the variable "positive DU in the past" was the only one to predict the presence of triatomines.This situation may be explained by the fact that insecticides applied in annexes are more exposed to sunlight, wind, and rain, which may reduce the duration of efficacy 27 .
Presence of Tr. sordida in the municipality of Dr. Camargo may be explained by the variables "positive DU in the past", "uninhabited DUs", "empty houses", "chicken coops", "pigsties", "number of annexes", and "proximity to scrub forest", either singly or in combination.DU annexes in this municipality were concentrated in a peridomiciliary area of approximately 200m 2 , located on average 275m from modified woodlands.Diotaiuti et al. 3 reported similar conditions for presence of Tr. sordida in the peridomiciliary environment.
In the Municipality of Porto Rico, the analyzed variables did not statistically explain the presence of DUs infested by P. megistus as compared to R. neglectus.However, this statistical result was expected, since this municipality is not a priority for Chagas disease control.Therefore, such infestations in Porto Rico could be incidental.
This study focused on current triatomine infestation in the rural areas of certain municipalities in the State of Paraná.In order to pre-Cad.Saúde Pública, Rio de Janeiro, 20(5):1191-1200, set-out, 2004 vent such situations in the future, the health authorities responsible for epidemiological surveillance in these municipalities, together with the residents themselves, must be aware of (and informed about) their responsibility in maintaining the commendable control of Chagas disease that has been achieved in Brazil as a whole.

Contributors
A. L. Falavigna-Guilherme contributed to the elaboration of the research project, literature review, data collection and analysis, and drafting of the final article.R. Santana collaborated in the data analysis.E. S. Lorosa and G. C. Pavanelli participated in the data analysis.S. M. Araújo contributed to the data analysis and revision of the final article.

Figure 1 Location
Figure 1 Location of municipalities in the State of Paraná, Brazil, where study was performed from June 1996 to February 2000.
, using T. cruzi Biolab antigens and Sigma conjugate.Positive results were later confirmed by IIF in serum derived from a new blood sample collected via venipuncture.Seropositive subjects (titer ≥ 1:40) were referred to the Health Care and Comprehensive Education Program for Chagas Patients (ACHEI) for the Municipality of Maringá and northwest region of Paraná, a follow-up program for Chagas patients run by UEM

Table 1
Investigation of triatomine infestation in domiciliary units (DUs) in rural areas of municipalities in Paraná State, Brazil, from June 1996 to February 2000.

Table 2
Inhabited and uninhabited domiciliary units (DUs) investigated for triatomines in Paraná State, Brazil, from June 1996 to February 2000.

Table 3
Prevalence of T. cruzi antibody in humans and domestic animals in rural areas of municipalities in Paraná State, Brazil.