Influence of climatic variables on the epidemiological situation of dengue in the Ceará- Brazil

Dengue is a worldwide public health problem; studies show a correlation between dengue and climatic variables. Thus, this research aimed to investigate the correlation of the number of cases of patients with primary dengue symptoms (NCPPDS) with the compensated average temperature (TCA), average precipitation (PA) and compensated average relative humidity (RHCA) for the state of Ceará (2013-2018), in addition to identifying the municipalities most affected by the disease. For this analysis, institutional databases were collected and the data were compiled and processed through Statistical Package for the Social Science software. The association between climatic variables and NCPPDS was made using Pearson's correlation. It was noted that the city of Fortaleza was the municipality most affected by the disease, followed by its metropolitan region. Pearson's correlation was significant and inversely proportional between the NCPPDS and TCA in the years studied. In the NCPPDS analysis and PA, a directly proportional significant correlation was observed in the years 2013, 2017 and 2018. Regarding RHCA during the years 2013, 2015, 2017 and 2018, a directly proportional significant correlation to dengue cases was also observed. These results showed that precipitation and humidity directly influenced the number of dengue cases in the state of Ceará, and in accordance with studies in other Brazilian regions, these findings represent a general picture for dengue spreading.


Introduction
Dengue is a viral disease transmitted by arthropods, which presents four serotypes of dengue virus (DENV-1, DENV-2, DENV-3 and DENV-4), transmitted by Aedes spp. mosquitoes. Dengue has evolved from a casual disease to a major public health problem with considerable social and economic effects, due to the increasing in its geographical amplitude, number of cases and the severity of the disease (Guzman & Harris, 2015). The number of dengue cases is estimated at 390 million infections per year and the Americas contribute 14% of these infections, more than half of which occurred in Brazil and Mexico (Bhatt et al., 2013). Global warming is believed to induce a gradual climate change, then probably tropical insects might expand their habitats (Zell, Krumbholz & Wutzler, 2008).
In 2012, dengue was considered the most important viral disease in the world, and it is found mainly in tropical and subtropical regions of the planet, where occurs preferably in urban and semi-urban areas, with risk factors influenced by local spatial variations of precipitation, temperature, relative humidity, degree of urbanization and quality of vector control services.
Severe dengue is one of the leading causes of acute diseases and deaths among children in some countries in Asia and Latin America (WHO., 2012. Brazil has 26 states and the Federal District, and an estimated population of 210 million people (IBGE., 2019). In 2018, the country registered 265,934 probable dengue cases, with an incidence rate of 127.5 cases/100,000 inhabitants, of which 174,724 (65.75%) were confirmed. The Northeast region of Brazil in recent years has shown an expressive number of probable dengue cases when compared to other regions (Brasil., 2019). The state of Ceará located in the Northeast of the country has an estimated population of 9 million people (IBGE., 2019), being composed of 184 municipalities, the predominant climate in the state is semiarid hot tropical (IPCE., 2017), and the historical series of dengue in Ceará reveals its epidemiological behavior and justifies the classification made by the Ministry of Health as a "vulnerability area with very high risk" for the occurrence of this disease (Lima et al., 2014).
Some studies indicate that the incidence of dengue cases has a strong correlation with climatic variables (temperature, rainfall and air humidity, among others), since these climatic conditions favor the increase of breeding grounds for the transmitting mosquito and facilitate the development of the vector (Mutheneni et al., 2017;Viana & Ignotti, 2013). Vectorborne diseases are among those most affected by climate change (Flahault, De Castaneda & Bolon, 2016). Recent research has used computational and mathematical models to make predictions and study the effects of climatic variables on the incidence of dengue in different places in the world. Mordecai et al. (2017) used mechanistic transmission models to derive predictions of how the probability and magnitude of Zika, chikungunya and dengue transmission are affected by changes in temperature, the values presented were compatible with real data. Wu et al. (2018) presented the non-linearity of the incidence of dengue in relation to average temperature and relative humidity in Guangzhou-China, suggesting that there are thresholds in these effects, below which there are significant positive effects.
Using a linear and non-linear model Xu et al. (2019) quantified the effects of average monthly temperature and relative humidity in cases of severe dengue in Thailand, in some provinces studied the risk of occurrence of severe cases of dengue increased with the increase in average temperature, in other regions this result was not predictable, as well as the relative humidity that significantly affected the occurrence of severe cases of dengue in Northeast and Central Thailand, with optimal intervals observed for each region.
Given the need for studies on dengue, since this is a worldwide problem, this research aimed to investigate the correlation of the number of cases of patients with primary symptoms of dengue (NCPPDS) with compensated average temperature (TCA), average precipitation (PA) and compensated average relative humidity (RHCA) for the state of Ceará (2013)(2014)(2015)(2016)(2017)(2018), in addition to identifying the municipalities most affected by the disease.

Methodology
This research was a quantitative-descriptive study of the variables that can influence the epidemiological situation of dengue in the state of Ceará during the period 2013-2018, by analyzing the frequency of incidences and statistical correlations, according to the methodology proposed by Chizzotti (2010).

Study location
The state of Ceará has an area of 148,894,757 km 2 and is located in the Northeast Region of Brazil between the coordinates of latitude 7°21'32"S and longitude 39°02'44" W ( Fig.1), with its limits to the West Piauí, to the East Rio Grande do Norte and Paraíba, South Pernambuco and North Atlantic Ocean. The state of Ceará has a strategic location for international tourism due to its proximity to Europe, North America and the African Continent, presents 184 municipalities, with the city of  The predominant climate in the region according to the Köppen-Geiger climate classification is semi-arid (BS'w'h') in 80% of the territory, the remaining 20% are almost entirely classified as sub-humid climate (Aw') present in the coastal region (Andrade et al., 2012).

Data collection
Data on the epidemiological situation of dengue in the state of Ceará (2013Ceará ( -2018 were obtained from the database of the Ceará State Health Secretariat through the Notifiable Diseases Information System (SINAN), and information was collected on the NCPPDS and the municipalities most affected. TCA and RHCA data for the analyzed period were extracted from eleven conventional meteorological stations of the National Institute of Meteorology (INMET) of Brazil, the stations are located in different regions of the state of Ceará (Table 1).

Data analysis
The data were analyzed using the software Statistical Package for the Social Science SPPSS version 23 and Microsoft Excel. Pearson's correlation given by equation ( Research, Society andDevelopment, v. 10, n. 12, e181101220313, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i12.20313 5 The monthly compensated average temperature was calculated by equation (II), and the relative humidity of the air, which also uses the value of the compensated average, measured by equation (

Epidemiological situation of dengue
The analysis of the databases obtained from SINAN (2013SINAN ( -2018 showed that the NCPPDS in the state of Ceará ( Figure   2) occurred more frequently in April, May and June. It was also noted that the year 2015 had the highest number of cases in this quarter. All NCPPDS was confirmed as dengue, severe dengue or dengue with alarm signs using laboratory and clinicalepidemiological criteria. In the series of data analyzed 2015 was the most incident year in NCPPDS, presenting about 57,261 cases in the state of Ceará. Source: Authors.
Since 1986, there have been reported cases of dengue in the state of Ceará and the four serotypes (DENV 1-4) were characterized. The disease manifested itself in an endemic form in the state with a record of seven major epidemics in the years 1987, 1994, 2001, 2008, 2011, 2012 and 2015. In 2015, the autochthonous transmission of the chikungunya and Zika virus was confirmed in the state (Brasil., 2017).

Municipalities most affected by dengue
During the period analyzed with data from SINAN (2013SINAN ( -2018, it can be noted that dengue was present in almost all municipalities in Ceará, with a higher incidence in more urbanized areas. Cases reported as dengue appears on a smaller scale in the North/East/West coast of the state as well as in mountainous cities. In the six years investigated in this study, the city of Fortaleza, capital of the state of Ceará, presented the largest number of confirmed cases of dengue ( Table 2). Municipalities that are part of the metropolitan region of Fortaleza also presented a significant incidence of confirmed dengue cases in the period studied. The city of Fortaleza occupied the first place in the ranking of the cities most affected by dengue (2013)(2014)(2015)(2016)(2017)(2018), the municipality has about 8,000 inhabitants per square kilometer, the highest population density among Brazilian capitals. The city has a semi-arid climate and an average minimum temperature of 24°C, the average maximum temperature is 30°C, its climatic conditions are influenced by the Intertropical Convergence Zone (ZCIT), by seasonal phenomena of El Niño, which occur in the dry season and La Niña, marked by rains above the average regional multiannual rate (Paula et al., 2013). influence on the population of mosquitoes, which generally inhabit peridomicile environments, a fact that explains the highest averages of arboviruses in the metropolitan region, due to population agglomerates and favorable tropical climate, with humidity and precipitation, which favor the presence of the vector, and consequently a greater number of cases.
The proximity to low-income urban and peri-urban centers was also associated with an increased risk of dengue, indicating that the human movement between population centers is an important facilitator of the spread of the disease (Bhatt., 2013).

Average compensated temperature (TCA)
The annual TCA data indicated as a result of the series studied for the state of Ceará (2013)(2014)(2015)(2016)(2017)(2018), an average between 26.23-27.23°C. It was verified through the analysis of normality with the Shapiro-Wilk test (Field., 2009) that the values presented are normal, presenting only small deviations of normality (Table 3). Most of the Northeast region has an average annual temperature of around 26°C, nevertheless there are places with higher values (Correia et al., 2017). In the study on climatic scenarios for the Northeast region of Brazil Santos et al. (2010) analyzed data from 89 meteorological stations , in this analysis it was observed that the average air temperature  Normally, rainfall in the state of Ceará has high spatial and temporal variability throughout the year, months from February to May are known as the rainy season, when it is observed about 70% of the annual precipitation. According to the climatology from 1981 to 2010, the PA in the rainy season in the state is 600.7 millimeters (Cortez, Lima & Sakamoto, 2017).
In this study, the PA in the rainy season (2013-2018) varied from 364.3-600.5 mm, with the last two years having the highest rainfall indexes 551.6 and 600.5 mm. The rainfall levels found are compatible with other semi-arid regions of the country (Lucena, Cabral Junior & Steinke, 2018;Francisco & Santos, 2017;Silva et al., 2018).

Compensated average relative humidity (RHCA)
The

Bivariate correlation between NCPPDS and climate variables
In the analysis of the correlation between the NCPPDS in the state of Ceará (2013-2018) and the climatic variables (TCA, PA, and RHCA), it was possible to observe a strong degree of relationship between the variables in the period investigated (Table 4).  (Evans., 1996).
Based on the analysis of the results shown in Table 4 and Figure 2, it was confirmed that the incidence of NCPPDS in  Table 4 showed variance shared between NCPPDS and TCA in 2013, 2015, at the level of 69%, 81%, 40%, 41%, and 58%, respectively.
The influence of meteorological variables on the prevalence of diseases transmitted by Aedes aegypti in Fortaleza indicating an inversely proportional relationship. It is also reported that the average air temperature has low monthly thermal amplitudes, with the minimum and maximum monthly thermal averages being optimal for the proliferation of the dengue vector, thus precipitation becomes the differential in the climatic dynamics of the disease (Magalhães & Zanella, 2013).
In the state of Paraíba, also located in northeastern Brazil, the analysis of the spatial modeling of dengue cases and climatic variables for three cities in that state during the period (2011)(2012)(2013)(2014) showed that the average temperature, maximum temperature, and minimum temperature did not present good correlation with dengue cases, the authors point out that the low seasonal thermal amplitude of these climatic variables in the study area may have implications for this result, in this way, Ae.
aegypti always find ideal temperatures for its development throughout the year (Almeida & Silva, 2017). Hay et al. (2000) through the analysis of monthly time series of severe dengue in Bangok, with monthly averages of temperature and precipitation collected during 33 years, showed that the interannual periodicity of dengue was not accompanied by similar periodic cycles of temperature and precipitation. The authors concluded that intrinsic factors such as population immunity had a greater impact than the climate behind the epidemics.
Some studies cite temperature as one of the climatic variables strongly correlated with dengue, these studies point out implications of temperature in all stages of development of the mosquito transmits dengue, as well as the interactions in the different seasons of the year (Massad et al., 2011;Mutheneni et al. 2017;Vezzani, Velásquez & Schweigmann, 2004).
In contrast, in this research there was no positive correlation between the NCPPDS and annual TCA for the state of Other studies in some regions of Brazil also showed a positive correlation between precipitation and the frequency of dengue cases, a higher number of dengue cases was found in periods with greater rainfall intensity (Duarte et al., 2019., De Jesus Corrêa, Da Costa & Pereira, 2016, Siqueira et al., 2018Silva & Magalhães, 2017). In particular, for Fortaleza, the correlation analysis of the prevalence of diseases transmitted by the mosquito Ae. aegypti and the influence of climatic variables (2013-2017) agrees with the findings of this research, where it was evidenced a strong relationship between dengue and precipitation and relative humidity (Araújo, Uchoa & Alves, 2019). Contrary to these results, there are studies that report the absence or weak correlation between the number of dengue cases and precipitation (Barbosa & Da Silva., 2016., Câmara et al., 2009. Two major dengue epidemics were recognized in Puerto Rico in the years 2007 and 2010, the analysis of the correlation between dengue cases and precipitation showed a greater number of dengue cases in coastal areas, especially in the north of the coast, places on the island that received more annual rainfall had a higher incidence of dengue, rain levels may have contributed to these epidemics (Schimer et al., 2011).
In Delhi, capital of India, during the period from 2011 to 2016, Bisht et al. (2019) revealed that rain peaks and dengue cases showed a two-month difference in the latency period, when this period was adjusted, a high positive correlation was noted that demonstrates a clear dependence of dengue cases on precipitation.
The contribution of rain patterns to the increase in dengue epidemics was also noted in Sri Lanka from 2004 to 2015, the country experienced progressive growth of the disease, mainly in urbanized areas and with a high population density. In the Colombo district, there was a great spread of the disease and the precipitation factor was shown to play a vital role in the spread of dengue, thus, a strong relationship was noted between the number of dengue cases and precipitation. It can be observed that generally the dengue epidemic is activated significantly after 3-4 weeks of the beginning of the rains (Edirisinghe, 2017).
Pearson's correlation discussed in this research identified a significant relationship between NCPPDS and PA for three years. In all series studied there was a higher incidence of dengue in the rainy months, a fact that can be correlated with precipitation. Although, as reported in the literature, the state of Ceará presents high climatic variability and spatial variability of precipitation (Martins & Vasconcelos Júnior, 2017). correlation. The months of the rainy season had high humidity and higher NCPPDS ( Fig. 2 and Fig.4), which corroborates with the result exposed in Pearson's correlation matrix (Tab.4). The analysis of variance of the NCPPDS and RHCA showed shared variance in 2013, 2015, 2017, and 2018, at the level of 56%, 48%, 79%, and 61%, respectively. Rao et al. (2018) reported the prevalence of dengue in India from the month of July with a peak in September, and it is notorious that climatic conditions of rain, temperature, and relative humidity favor the proliferation of the disease. Continuous rains for a long period cause a temperature reduction from June to November, favoring the increase in relative humidity and decreasing the evaporation rate, thus maintaining secondary reservoirs containing rainwater, which makes it favorable for the development of the dengue vector.

NCPPDS in
In Tawau, Malaysia, dengue prediction models based on 12 years of study made a statistically significant crosscorrelation between climate predictors and the disease, in this data analysis an average relative humidity of 82.5% was verified and the fact that all epidemics of dengue occurred with average relative humidity above 80% (Jayaraj et al., 2019).
Humidity is reported by Vargas et al. (2010)  In this research, it was possible to notice the interrelationship between the climatic variables, hotter and drier months tend to have lower levels of precipitation and humidity, colder and humid months tend to have higher values of precipitation and humidity, this fact may favor the increase breeding sites of the dengue-transmitting mosquito, as a consequence there may be an increase in the NCPPDS. The relationship between dengue and the climatic variables studied for the state of Ceará (2013)(2014)(2015)(2016)(2017)(2018) can be summarized in Figure 5, which was built based on the analysis of Pearson's correlation shown in Table 4. Source: Authors.
It was observed a significant correlation between dengue and climatic variables, and a directly proportional relationship between dengue ( Figure 5), relative humidity and precipitation (↑↑), and inversely proportional between dengue and temperature (↑↓).
Based on the data discussed in this research and other studies on dengue, it is clear that dengue is influenced by climatic variables, although other factors may collaborate to trigger major epidemics of the disease, such as lack of urban planning and health education.

Conclusion
The analysis of the dengue databases in the state of Ceará (2013-2018) obtained by SINAN together with the climatological data of TCA, PA, and RHCA, coming from INMET and FUNCEME allowed to identify a significant correlation between climatic variables and NCPPDS. In all the annual series analyzed, the city of Fortaleza, capital of the state of Ceará, was in the ranking of the cities most affected by dengue, followed by its metropolitan region, and 2015 was the year with the highest incidence of dengue cases in the state.
Pearson's correlation showed a strong and inversely proportional relationship between NCPPDS and TCA in five of the years studied. PA and RHCA have a directly proportional relationship with the NCPPDS. In four years of the series studied RHCA had a significant correlation with the number of dengue cases, the variable with the greatest direct and positive association, followed by precipitation.
All climatic variables analyzed in this research are associated with the NCPPDS. The degree of association and how it occurs differs like the relationship. The inversely proportional correlation of TCA with dengue does not reveal that this variable is not relevant for situational analysis of the disease, since the mechanisms of dengue transmission, the stages of development of the dengue-transmitting mosquito are notably affected by variations in temperature. Factor as high climatic variability of the state of Ceará may have caused the inverse nature of this relationship, which does not mean that the temperature in the state is not an associated factor with the NCPPDS, there is no simultaneity in these values, that is, high-temperature values do not cause necessarily increase in cases of dengue, since the semi-arid climate would be conducive to the proliferation of mosquitoes and consequently the disease all year round.
These results can be useful for developing strategies to combat dengue since it was clear that rainy months in the state of Ceará have a higher incidence of cases of the disease associated with RHCA between 78.20 to 87.35%, these climatic variables can be used as good predictors of disease infestation. Therefore, the differential in this association is to perceive the interrelationship between the variables, correlated climatic characteristics can favor the generation of dengue epidemics worldwide, the complexity of this disease requires that environmental and social factors are also observed so that it is possible to expand dengue prevention and control measures. Therefore, the differential in this association is to perceive the interrelationship between variables, correlated climatic characteristics can favor the generation of dengue epidemics worldwide, the complexity of this disease requires that they also be observed environmental and social factors, so that it is possible to expand the prevention and control measures of dengue.
Future research and work may investigate the socioeconomic situation, age and sex of people most affected by dengue in the state of Ceará, a temporal analysis or prediction of epidemic events through other statistical models may increase the level of correlation of climate variables with the dengue.