Analysis of sandflies ( Diptera : Psychodidae ) in Barra do Garças , State of Mato Grosso , Brazil , and the influence of environmental variables on the vector density of Lutzomyia longipalpis ( Lutz & Neiva , 1912 )

Introduction: Leishmaniasis is an infectious and parasitic zoonotic, non-contagious, vectorborne disease caused by protozoa of the genus Leishmania. In Brazil, the major vector of Leishmania (Leishmania) infantum chagasi (Cunha & Chagas, 1934) is Lutzomyia longipalpis. Barra do Garças, State of Mato Grosso, was designated as a priority area by the Brazilian Ministry of Health for american visceral leishmaniasis, and it is important to identify the vector species present in this municipality. Our objective was to raise sandflies and study the influence of environmental variables on the vector density of Lutzomyia longipalpis. Methods: We performed entomological monitoring in 3 districts using Centers for Disease Control and Prevention (CDC) light traps and recorded human cases of american visceral leishmaniasis in the city. We calculated the relative frequency and richness of sandflies and applied a transfer function model to the vector density correlate with relative humidity. Results: The sandfly population was composed of 2 genera and 27 species, totaling 8,097 individuals. Monitoring identified Lutzomyia longipalpis (44%), followed by Lutzomyia lenti (18.9%), Lutzomyia whitmani (13.9%), Lutzomyia carmelinoi (9.1%), Lutzomyia evandroi (5.1%), Lutzomyia termitophila (3.3%), Lutzomyia sordellii (1.9%), and 20 other species (<4%). The male:female ratio was 3.5:1. We observed high species diversity (Dα = 6.65). Lutzomyia longipalpis showed occurrence peaks during the rainy season; there was a temporal correlation with humidity, but not with frequency or temperature. Conclusions: The presence of Lutzomyia longipalpis in the urban area of Barra do Garças underscores the changing disease profile, which was previously restricted to the wild environment.

According to estimates by the World Health Organization, leishmaniasis is prevalent in 4 continents and endemic in 88 countries 1 .AVL is endemic in Brazil, reaching 5 regions, with the occurrence of human cases in 24 states.
In 1913, AVL was first described in Brazil from an autopsy of a patient from Boa Esperança, State of Mato Grosso 2 .The disease has rapidly expanded into the interior of Mato Grosso (MT) following the process of uncontrolled urban expansion, coupled with the intense migratory flow between the towns of the central-south with the north and southeast of the State 3 .
From 2001-2006, the municipality of Barra do Garças-MT recorded 19 human cases of AVL, 2 of whom progressed to death 4 .The municipality is classified as a priority area for AVL, alternating between areas of moderate and/or sporadic transmission 5 .The presence of L. longipalpis and L. cruzi in the municipality of Barra do Garças-MT was reported between 1996 and 2004 6 .Seropositive dogs have also been reported 3 with average positive rates of 12%.
Given the above, and the increasing expansion of AVL in Mato Grosso, it became crucial to monitor the sandfly species present around residences in Barra do Garças-MT.
The objective of this study was to monitor the sandflies in neighborhoods of the City of Barra do

RESULTS
Queiroz MFM et al -Sandflies in Barra do Garças, MT and influence environmental on vector density Garças-MT with a record of human cases of AVL.We also studied the influence of environmental variables, i.e., temperature, relative humidity, and rainfall, on the vector density of L. longipalpis.
Barra do Garças-MT has a population of 56,423 inhabitants and a land area of 9,079 km 2,7 , and is located along the Araguaia and Garças rivers (Figure 1), forming an important tourist resort, especially during June-August by offering extensive freshwater beaches that are visited by tourists from Mato Grosso, Goiás, Distrito Federal, among others.It is situated at latitude 15°89'00" and longitude 52º25'66" and is located at 318 m above sea level 8 .Its climate is Aw (hot and humid), and the dry season is from May-October and the rainy season is from November-April 9 .The average annual rainfall is 1,750mm and the mean annual temperature is 24°C (minimum 0°C, maximum 40°C) 10 .The predominant vegetation type is savanna, with small portions of forest and a transition zone 6 .The collection of sandflies was carried out monthly in the neighborhoods of Novo Horizonte, São José, and Vila Maria.These locations were chosen because of the presence of human cases of AVL, and these are peripheral areas, located southwest of Barra do Garças-MT.Samples were collected for 4 consecutive nights between 17:00-07:00 for 24 months from September 2004 to August 2006, thereby covering different seasons.
The homes were chosen to maintain favorable environmental characteristics for the life cycle of sandflies, as there were organic matter in soil (leaves, fruit, rotting logs, and feces/manure), abundant vegetation (fruit trees), livestock (dogs and chickens), and market gardens.We measured the following variables: temperature (°C), relative humidity (%), rainfall (mm), and geographic coordinates.The monthly rainfall averages for the period studied were obtained from the Brazilian Company of Airport Infrastructure from the station based in Barra do Garças-MT.
The trapped insects were killed with ethyl acetate, sorted, and stored in plastic vials containing 70% alcohol for further clarification and identification at the species level by the Laboratory of Entomology, Central Level of the State Secretariat of Health of Mato Grosso (SES/MT), using phase contrast optical microscopy and a taxonomic key 12 .From February 2005, we also identified the sex of the collected individuals.
The relative frequency of the sandflies was calculated using percentages, and the Margalef 's index (Dα) was used to calculate wealth, translated from faunal composition 13 .
The influence of temperature and relative humidity on the number of sandflies was assessed using the transfer function model 14 .We also used analysis of variance (ANOVA) to determine whether significant differences occurred between the vector density of L. longipalpis and rainfall in different seasons -winter and summer -in the years studied.To check whether there was a correlation between relative humidity and rainfall, we applied the Bartlett correlation test 14 , and this was tested for up to a 2-month lag.
In the neighborhoods of the City of Barra do Garças-MT in the period from September 2004 to August 2006, the sandfly population was composed of 2 genera and 27 species, totaling 8,097 sandflies collected from peridomestic locations; of these, 44% were L. longipalpis.From February 2005, when we began to sex the species, the proportion of males to females of L. longipalpis was approximately 3.5:1 (Table 1).
The index of species richness (Dα) was 6.65, where Dα values > 5 indicate high diversity.L. longipalpis was the most abundant species, being collected in all months surveyed (Figure 2).
In the period studied, the temperature varied from 16-40°C, and the relative humidity ranged from 24-92% while maintaining the standard of Mato Grosso, Brazil.
All cross-correlations between the frequency of L. longipalpis with temperature at different time periods were statistically zero (p > 0.05).However, there was a significant correlation between the frequency of L. longipalpis with the relative humidity in the same (p < 0.07) and previous (p < 0.05) month.
How had a significant correlation between the relative frequency of L. longipalpis and humidity can apply a transfer function model for a time.To better explain the model, the monthly relative humidity (X t ), the monthly frequency of L. longipalpis (Y t ), and the unit of time (t) for each month are considered.
The model transfer function that relates these 2 series is given by: Y t = 1.17Y t-1 -0.31Y t-2 + 2.86 X t -1.14X t-1 -0.765a t-1 + a t , indicating that the monthly frequency of L. longipalpis (Y t ) depends on the frequency of L. longipalpis in the previous month (Y t-1 ) and in the second preceding month (Y t-2 ) and on the relative humidity in the same month (X t ) and in the previous month (X t-1 ).The corresponding weights are given by the numerical values of the model.Moreover, the same monthly frequency of L. longipalpis (Y t ) also depends on other random factors, and the same month and the previous month are represented by a t and a t-1 , respectively.The association between the frequency of L. longipalpis captured and the variation of rainfall is represented in Figure 2. The annual rainfall was 1,668.9mm, 1,489.6 mm, and 1,406.3mm for the years 2004, 2005, and 2006, respectively.
ANOVA applied to verify if the seasons, i.e., dry and wet, interfere with the vector density of L. longipalpis provided a p value of 0.00058 (p < 0.05), indicating that there were significant differences for various periods, and the vector density of L. longipalpis was highest in the rainy season from November 2004 to March 2005.
When testing the influence of rainfall on relative humidity, there was a significant correlation between rainfall with relative humidity in the same month (r = 0.72, p < 0.05) and in the subsequent 2 months (r = 0.60, r = 0.50, p < 0.05).
The third most abundant species, L. whitmani, is the main vector of american tegumentary leishmaniasis (ATL) in the State of Mato Grosso and in Barra do Garças-MT, being found in peridomestic and indoor areas (V Correa, GBM Lima, MFM Queiroz: unpublished data).This species is in the process of domestication and is widely distributed with greater adaptation to man-made changes to the environment 15 .
Despite the low density of L. flaviscutellata and L. wellcomei in the urban area, their presence is of concern given the importance they represent to public health.L. flaviscutellata is a common species of forest and halfdomestic environments.Since they are vectors of Leishmania and arboviruses, their preferred habitats are chicken coops, pig pens, corrals, and other attachments of domestic animals and the external and internal walls of human homes 15 .
L. wellcomei, another vector for ATL, has been reported for the first time in Mato Grosso.The presence of this species Peak capture and rainfall did not coincide for the 2 years studied, suggesting that rainfall encourages the growth of the sandfly population, but not excessively so, as reported by other authors 25,26 .The direct effect of large amounts of rain can change the biological cycle of the sandfly.It is possible that excessive rainfall alters the availability of the organic matter that they use for the deposition of eggs and for their larvae to feed on; however, it could also kill the immature flies.Therefore, there seems to be a threshold level of rainfall that is favorable for the reproductive cycle of sandflies.
In the first year of entomological monitoring (September 2004-August 2005), we observed a large number of L. longipalpis in the study area; however, this was followed by a decline in the second year.This fact can also be associated with environmental management (e.g., tree pruning, removal of organic matter, and eliminating some chickens and pigs) that was conducted in the peridomestic environment, as confirmed by other authors 5,33 .
The presence of L. cruzi in the urban area was previously reported in Barra do Garças-MT 6 , but this species was not collected during the 2 years of monitoring.The occurrence of a species in a local community is partially determined by its adaptation to the conditions and available resources and partly by competitive interactions with other species 34 ; as these species are very close phylogenetically, both must use the same ecological resources, and competition must occur 18 .Therefore, interspecific competition could explain the absence of L. cruzi during this period.
The results indicated a worrying situation for the urban area of Barra do Garças-MT, as L. longipalpis was the most frequent and abundant species, being collected in every month of the year and found in all households surveyed, since it is the main vector of Leishmania (Leishmania) infantum chagasi in Brazil 19,35 .Entomological knowledge about the seasonality and population dynamics of L. longipalpis in Barra do Garças-MT allows the targeting of prevention and control activities for AVL, optimizing financial resources and actions, and thus promoting the strengthening of public health.
The importance of detecting L. longipalpis is the fact that this species is more adapted to the epidemiology of the disease, is widely distributed throughout the country, and is present in areas where AVL is a constant concern of public health management programs 36 .
The high richness of sandflies found reinforces the importance of further studies in order to raise more qualitative and quantitative information on the transmission cycle of leishmaniasis in the municipality of Barra do Garças-MT.
We thank the municipal health team for logistical and technical support, especially to the Agents to Combat Endemic Diseases, on behalf of Elis Tagino de Lima and Andrea de Menezes, who collaborated with the collections, at Central Level/SES/MT, and also for logistical and technical support, especially in the identification of the species collected; and the other technicians of the Regional Health Barra do Garças who directly or indirectly contributed to the creation of this monitoring program.

FIGURE 1 -
FIGURE 1 -Geographical location of the municipality of Barra do Garças, State of Mato Grosso, Brazil.

FIGURE 2 -
FIGURE 2 -Relationship between the vector density of Lutzomyia longipalpis collected with CDC type light traps and the average rainfall from September 2004 to August 2006 in the town of Barra do Garças-MT.