Profile of American tegumentary leishmaniasis in transmission areas in the state of Minas Gerais, Brazil, from 2007 to 2017

Background American tegumentary leishmaniasis (ATL) is a widespread anthropozoonosis caused by protozoa of the genus Leishmania and is considered a serious public health problem. The aim of this study was to provide a descriptive analysis of confirmed ATL cases and evaluate the spatial distribution of ATL in high-risk transmission areas from the state of Minas Gerais, Brazil. Methods An ecological, analytical, and retrospective study of the confirmed cases of ATL in Minas Gerais from 2007 to 2017 was conducted. To characterize these cases, multiple correspondence analysis and georeferencing of the ATL prevalence rates in the municipalities were conducted based on variables obtained at Sistema Nacional de Agravos de Notificação and Instituto Brasileiro de Geografia e Estatística databases. Results There were 13,025 confirmed cases of ATL from 74.4% (635) municipalities of Minas Gerais, corresponding to a prevalence rate of 66.5 cases for every 100,000 inhabitants. Males aged 20 to 59 years and individuals who attended elementary school were most affected with ATL. Multiple correspondence analysis presented an accumulated qui-squared value of 44.74%, proving that there was a relationship between the variables, including ethnicity, age, pregnancy status, zone of infection, and number of cases. Conclusion We confirmed that ATL is endemic to Minas Gerais, and there is high risk of infection within the municipalities due to a high rate of parasite transmission. The occurrence of infection in children, pregnant women, and the indigenous population demonstrates the need for the government to expand social policies aimed at vulnerable groups.


Background
American tegumentary leishmaniasis (ATL) is a widespread zoonotic disease that has a global impact [1]. It is caused by the protozoa of the genus Leishmania and is considered a serious public health problem [2]. Phlebotomine sand flies, such as Nyssomyia spp. and Lutzomyia spp., belonging to the family Psychodidae (subfamily Phlebotominae), act as vectors for Leishmania in the Americas and play an important role in disease dissemination among humans and domestic animal reservoirs [3][4][5]. In Brazil and South America, the major agent causing ATL is the Leishmania (Viannia) braziliensis [6].
Recently, ATL has been reported in all Brazilian states, and its epidemiological pattern has gone through changes regarding its transmission, with the occurrence of peaks every five years [7]. From 2007 to 2017, 232,989 cases of ATL were reported in Brazil, with a mean prevalence of 118.39 cases annually for every 100,000 inhabitants.
In the state of Minas Gerais, ATL was first reported in 1940, and its transmission has been reported from both rural and peri-urban areas [8]. Thus, the objective of the present study was to use descriptive analysis to characterize ATL cases from the state of Minas Gerais and demonstrate their spatial and temporal distribution from 2007 to 2017.

Development of the study
An ecological, analytical, and retrospective study of confirmed cases of ATL, which occurred between 2007 and 2017 in Minas Gerais, was conducted using the secondary database established by the Sistema de Informação de Agravos de Notificação (SINAN) and the Brazilian Institute of Geography and Statistics, also known as Instituto Brasileiro de Geografia e Estatística (IBGE).

Location of the study
The state of Minas Gerais is located in the southeastern region of Brazil and is divided into 853 municipalities. Its territorial area is 586,520.732 km 2 , and its vegetation is composed of 19.94% savannah, 10.33% Atlantic forest, and 3.48% caatinga; 33.8% of the total area is native forest [9,10]. According to the IBGE census conducted in 2010, Minas Gerais had a population of 19,597,330 inhabitants, corresponding to a population density of 33.41 inhabitants per km 2 [9].

Database
A total of 635 municipalities in Minas Gerais, which reported confirmed cases of ATL were considered as analytical units, and epidemiological aspects that could characterize these cases were evaluated using data obtained from SINAN. Demographic variables, such as gender, age, education level, zone of infection, and ethnicity were evaluated, in addition to variables, such as pregnancy status, annual and monthly frequency of disease, and ATL confirmation criteria. These variables were dichotomized or categorized to perform a multiple correspondence analysis. An adaptation of the criteria presented by [7,11] was used to classify the epidemiological condition of ATL cases from Minas Gerais, which considered the prevalence rate of each municipality positive for ATL. The number of ATL cases within each municipality was categorized as "low" for 1-20 cases, "moderate" for 21-50 cases, and "high" for 51-828 cases.

Statistical analyses
Multiple correspondence analysis of the data was performed using the software Stata 12.0 (Stata Statistical Software: Release 12. College Station, TX: StataCorp LP). Variables that comprised the correspondence model were selected using Pearson's chi-square test (P ≤ 0.05) and evaluated for possible associations between the number of cases and other variables using graphs, which were interpreted by evaluating the proximity of the variables under the category "number of cases." The intensity of association with values of cumulative inertia above 40% was considered [12].
For the epidemiological analysis of ATL from 2007 to 2017 in Minas Gerais, a control diagram was designed by calculating the mean, upper and lower limits, and standard deviation. By using this diagram, it was possible to evaluate the temporal evolution of the disease and detect possible alterations in its distribution pattern. The diagram data can be used to developing effective protective measures against the ATL [13].

Georeferencing
The prevalence rate of ATL in the municipalities, was calculated in a 95% confidence interval based on data obtained from SINAN, and a georeferencing map was constructed. The software QGIS 2.18.14 (QGIS Development Team, 2009, QGIS Geographic System, Open Source Geospatial Foundation) was used to map the spatial distribution of municipalities with prevalent infection using data for confirmed cases of ATL that occurred in Minas Gerais from 2007 to 2017 (Fig. 1). After the calculation of the prevalence rate, 20 municipalities were identified as having prevalent infection, with the highest number of cases observed in the state of Minas Gerais (Table 1).

Results
From 2007 to 2017, 13,025 confirmed cases of ATL from 74.4% (635) municipalities of Minas Gerais ( Fig. 1) were reported, corresponding to a prevalence rate of 66.5 cases for every 100,000 inhabitants. One third of these cases were reported from 15 municipalities evaluated in this study ( Table 1). The municipalities of São João das Missões, São João do Pacuí, São Domingos das Dores, Varzelândia, and Januária had the highest prevalence rates of ATL. Data for Belo Horizonte, Teófilo Otoni, Montes Claros, Ipatinga, and Patos de Minas were included in the variable analysis owing to their importance as economic centers of the state and because they reported a moderate to high frequency of the disease.
A total of 90.65% cases were diagnosed using clinicallaboratory criteria, while 9.35% of cases were diagnosed using clinical-epidemiological criteria. Laboratory diagnosis involved immunological examinations, such as the Montenegro intradermoreaction, parasitological tests with direct visualization of the parasite, in vitro isolation and culture, in vivo isolation, and polymerase chain reaction. The clinical-epidemiological criteria for diagnosing cutaneous leishmaniasis were applied only to individuals without access to laboratory diagnostic services and with  Subtitle: * per 100,000 inhabitants; Minas Gerais presented a prevalence rate of 66.5 cases for every 100,000 inhabitants residence, origin, or displacement to an area with confirmed cases. When considering gender and age variables, male individuals (60.21%) and adults aged 20 to 59 years (56.89%) were seen to be the most affected. The education-level variable showed that approximately 36% of the cases reported in the highlighted municipalities, were those of individuals who had elementary schooling. However, 50% of notification forms did not include this information in these locations. Upon evaluating the role of ethnicity in disease prevalence, we verified that the individuals most affected by ATL were brown skinned (44.05%). There were 24 cases of pregnant women with ATL, representing 0.59% of all forms of ATL notifications in high transmission areas ( Table 2).
The prevalence of confirmed cases of ATL is highlighted in Fig. 1. A cluster can be observed in north Minas Gerais, Rio Doce (east), and Jequitinhonha valley. The southern and Mineiro Triangle regions presented low to moderate prevalence.
The control diagram describes the expected ATL case average in Minas Gerais from 2007 to 2017 and suggests that Minas Gerais is an endemic location for the disease. Moreover, the average number of cases was slightly higher in the months when the temperature was warmer compared to the average number of cases during the colder months (Fig. 2). The correspondence analysis plot, demonstrating the epidemiologic profile of ATL in Minas Gerais, presented an accumulated chi-squared value of 44.74%, and the evaluated variables of the model have been highlighted by blue circles in Fig. 3. We observed that the number of case variables corresponded with variables, such as age, ethnicity, and education level, zone of infection, and pregnancy status. The associations among a set of variables, such as age, gender, ethnicity, zone of infection, and confirmation criteria was also verified.
An association among brown skin, age 4 (60 years or older), female sex, infection in rural areas, and diagnosis made based on clinical-epidemiological criteria was also analyzed.

Discussion
In the last 11 years, 13,025 cases of ATL were confirmed in Minas Gerais, and the highest prevalence rate was observed in municipalities from the north of the state and from Rio Doce (east) and Jequitinhonha Valley. These regions are known as zones of low development and poverty and present significant diversity among the phlebotomine species. These features may be related to the dissemination of ATL to humans and domestic animals, which act as reservoirs, in these areas. In addition, anthropogenic factors are also determinants of disease maintenance [14][15][16].
We verified that the most effective confirmation criteria were those from the clinic and the laboratory. Confirmation of ATL using the clinical-epidemiological criteria provides important epidemiological information, as it identifies the parasite and may help in taking control measures against it. This result is in agreement with the Ministry of Health Guidance for Health Surveillance (2017) [7], which claims that the parasitological method for ATL diagnosis may be conducted before the onset of treatment, especially in cases where the clinical evolution is not ordinary and/or there is a bad response to treatment.
Analysis of data on the age and gender of patients with ATL in Minas Gerais revealed that males of working age were more susceptible to the disease, a finding that corroborates with those of previous reports [17,18]. This might be because the occupations of males in this region, such as military occupation and construction, put them at a higher risk of being exposed to ATL vectors [19,20]. The low education level among most patients with ATL might be associated with a poor understanding of risk factors related to the disease. The occurrence of ignored notification forms this variable (50.25%) might make it difficult to implement educational measures targeted to specific risk groups according to the degree of education. Additionally, the absence of educational level data makes it impossible to evaluate this criterion and is a limitation of this study.
Concerning the ethnicity of the patient, the ATL infection pattern agrees with the racial-ethnic characteristics of the Brazilian population demonstrated by IBGE (2010) [11]. According to IBGE, there is a larger brown skinned population in Brazil compared to the populations of other ethnicities. However, São João das Missões, a municipality located 663 km from the capital Belo Horizonte also has a strong presence of indigenous populations. There are a total of 33 villages occupied by the Xakriabá in Belo Horizonte [21]. In these villages, 359 cases of ATL were reported over a period of 11 years; 333 (92.8%) occurred within the indigenous population. Moreover, this municipality presented the highest prevalence of ATL in Minas Gerais during the study period (3064.4 cases per 100,000 inhabitants). Thus, it is considered an area of high risk for the transmission of Leishmania ( Fig. 1 and Fig. 3). The data demonstrated greater susceptibility to Leishmania infection within the indigenous population. Furthermore, a previous study [22] had suggested that the high prevalence of ATL in this area was due to the presence of a large number of wild reservoirs and high rates of deforestation. The main species involved in this area was L. braziliensis, in agreement with data presented by a study conducted in 2017 [22,23], which reported on the presence of L. braziliensis in individuals from an indigenous village in Mato Grosso.
The MCA results showed that the peri-urban zone of infection was related to the municipalities that presented moderate to a high number of ATL cases. These results agree with those described in the manual of ATL surveillance [7] and with those of the survey conducted in 2017 [26], which considered the peri-urban zone an important infection area for the disease. Moreover, the literature suggests that the adaptation of the phlebotomine subfamily and anthropic activities that impact the environment are responsible for ATL transmission in the peri-domestic environment. For example, breeding animals close to households favored vector adaptation to the human domain and modified the ATL transmission cycle [27].
In the correspondence analysis, some associations between disease in pregnant women and children younger than 1-year-old and indigenous/yellow ethnicity was found. This might explain the higher susceptibility of these populations to ATL and can be linked to characteristics related to exposure to the vector and vector adaptation to the peri and/or intra-domestic areas [28].
The results observed for the associations among brown skin, age 4 (60 years or older), females, infection in rural areas, and disease confirmation using the clinicalepidemiological criteria are similar to those obtained in other published reports [2]. Furthermore, the present study verified that adult males are the most affected, while children and women are affected to a lesser degree.
Analysis of the control diagram showed that ATL is endemic to Minas Gerais, as the data on disease frequency in this area are similar to the results seen in the past in the absence of epidemics. In addition, we verified that during cold months there was a slight decrease in the occurrence of ATL cases, which demonstrated a correlation between temperature and phlebotomine activity, as described in a study conducted in 2017, which compared the activity of the vector during different seasons [29].
When taking the entire country into account, the average prevalence rate for ATL was approximately 118.9 cases for every 100,000 inhabitants, suggesting that the situation in Minas Gerais was less critical, as it has an average prevalence rate of 66.5 cases for every 100,000 inhabitants. However, the cutoff that determined whether certain locations were at a high risk, is 448.3 cases for every 100,000 inhabitants [11]. In this study, we found that the municipalities highlighted in Table 1 are at a higher risk of leishmaniasis transmission.

Conclusion
The present study provides a new perspective on ATL as it involved a multivariate analysis using multiple correspondence analysis and used epidemiological and demographic variable data openly available on SINAN and IBGE.
We verified that ATL was endemic to Minas Gerais, especially in municipalities with high transmission rates. The occurrence of the infection in children, pregnant women, and indigenous populations demonstrates the need for the government to implement social policies directed towards these vulnerable groups. Thus, it is essential to fully understand the epidemiology of ATL to implement control measures against this parasite.
ATL is a health priority in certain municipalities of Minas Gerais. Our study evaluated disease epidemiology and discussed the need for public policies to address ATL. Recently, Brazil decentralized its health services; such changes may worsen disease incidence and prevalence in this country, as the country has been struggling economically and lacks human and material resources required for disease control.
Cases of ATL notified through a secondary database can be a limitation for this type of epidemiological study as they can lead to an underestimation of disease prevalence, due to under-reporting, and cause collection bias. Moreover, ecological studies can be susceptible to ecological bias as they do not allow for analysis of individual data.

Acknowledgments
We thank the UFMG Veterinary School for making the epidemiology laboratory available for data analysis.
Authors' contributions MOKMP, TMO and ANAA conceived the study and drafted the manuscript; TMO, DSB carried the analysis and interpretation of data and drafted the manuscript; PEMP, SAD and MXS critically revised the manuscript for intellectual content. All authors read and approved the final manuscript.

Funding
We appreciate the financial support and scholarships provided by Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG) and by Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES). The financial support did not play any role in the design of the study or collection, analysis, and interpretation of data.