Evidence of Co-Circulation of Multiples Arbovirus Transmitted by Aedes Sp. Based on Laboratorial Syndromic Surveillance At Health Unit in Slum Area of Federal District, Brazil

Background: Vector-borne diseases, especially arbiviruses transmitted by Aedes sp. mosquitos, should be a health policy priority in Brazil. Despite this urgency, there are signicant limitations in the traditional surveillance system, such as lack of timely notication in identifying outbreaks at their onset and the systemic dismantling of entomological control in recent decades. Methods: Laboratory syndromic surveillance for acute febrile and/or exanthematous syndromes was developed at a health unit in a favela (Portuguese for slum) of the Federal District in order to identify the circulation of arboviruses transmitted by Aedes sp. mosquitos. Between June/2019 and March/2020, 131 valid participants were identied and tested by RT-PCR for dengue (by serotype), Zika and Chikungunya virus acute infection; and ELISA-IgM for dengue fever and Chikungunya 15-21 days after symptom onset, when the participant reported no respiratory signs (cough and/or coryza). The results obtained were compared with traditional epidemiological surveillance data for the study area. Results: At least 3 DENV-1 (2.3%), 4 DENV-2 (3%) and 1 CHIKV (0.7%) cases were conrmed in the laboratory, showing evidence of hyperendemicity even though laboratory syndromic surveillance had not reached the historic peak dengue fever months in the Federal District (April-May). When the results obtained here were compared with traditional epidemiological surveillance data, a signicant discrepancy was observed, including underreporting of Chikungunya infection. Conclusions: in addition to the risks posed to the study population, the area investigated and its respective socioenvironmental prole may be a potential site for spreading the virus, given the cosmopolitan presence of Aedes sp. and human mobility in the Federal District. It is also suggested that traditional epidemiological surveillance may be notifying different acute viral infections such as dengue fever, while underreporting other arboviruses transmitted by Aedes sp. mosquitos in the Federal District.


Background
The emergence and reemergence of neglected infectious diseases (hitherto de ned as tropical diseases) require continuous monitoring by the health agencies responsible for surveillance and intervention. This is especially important in Brazil, due to its continental size and large number of biomes under constant environmental disturbance, thereby promoting ecological imbalance. Among the main emerging infectious diseases, arboviruses transmitted by Aedes sp. mosquitos deserve the attention of health policies, as observed with the dengue virus in the Americas during the rst two decades of the 21 st century. The continent had the largest number of noti cations in the world, with Brazil contributing with the highest proportion and exhibiting an endemic-epidemic pattern every 3 to 5 years, in line with the prevalence of the serotype [1]. The country plays a key role in ampli cation and potentialization of mosquito-borne diseases, as stated by Bedin in 2007 [2], and demonstrated by the Aedes aegyptitransmitted Zika virus epidemic in 2015 which, despite its circulation in other countries, was only con rmed as having an association with congenital malformation syndrome after it emerged in Brazil [3].
With respect to the Aedes aegypti vector in Brazil, currently present in all the states of the country since its reemergence in the 1960s, the transmission of four dengue serotypes stand out: serotypes 1 and 4 in outbreaks during the 1980s, followed by serotype 2 in the 1990s, and serotype 3 in the 2000s [4]. It is also important to mention the Chikungunya virus, introduced in the country in 2013 [5], and the Zika virus in 2014 [3]. The species is also an urban vector for the yellow fever virus in Brazil, where repeated outbreaks were observed in recent years, associated with wild vectors (Hemagogus sp. and Sabethes sp.) in forest areas near urban zones [6]. Also important are other opportunistic arboviruses (mayaro, o'nyong-nyong, Oropouche, etc.), whose dynamics should be investigated [8,9] given the vector capacity of Aedes aegytpi [7]. Finally, Aedes albopictus, a secondary vector for the dengue virus in Brazil, is important in rural areas [10], but also observed in parks and green spaces in large cities [11].
This current hyperendemic scenario and the complex clinical ndings associated with arbovirus infections, such as neuroinvasive and Guillian-Barré syndromes, systematically overload the Brazilian health system, affecting the country's therapeutic and economic capacities [12; 13]. In addition, the congenital malformation syndrome associated with Zika virus infection promotes this agent to the group of STORCH infections (syphilis, toxoplasmosis, rubella, cytomegalovirus, and herpes simplex), thereby requiring prenatal monitoring [14]. The sympatric circulation of different serotypes of dengue and other arboviruses promotes interactions in the physiopathogeny and immunological response of the human host, resulting in severe outcomes such as hemorrhagic shocks and death [15]. The consequence of new arboviruses with similar dynamics to that of the dengue virus produces more complex transmission, resulting in a broad symptomatological spectrum while decreasing the sensitivity and speci city of the clinical-epidemiological criterion for diagnosis, underscoring the need for laboratory con rmation [16]. The high number of assymptomatic and oligosyntomatic infections [17], as well as false-negative diagnoses due to the low sensitivity of widely used methods [18], mean outbreaks related to the arboviruses may occur without the epidemiological surveillance system's detecting them in a timely fashion, thereby facilitating their spreading and epidemics [19]. Also considered a complicating factor in controlling mosquito-borne diseases is the systematic dismantling of vector surveillance in recent decades in Brazil, which contrasts with the country's history of this strategy, such as the erradication of Aedes aegypti in the late 1950s [20].
In the present study, laboratory syndromic surveillance (LSS) was conducted for the three arboviruses transmitted by Aedes sp. mosquitos in a slum of the Federal District of Brazil, for an expected 1-year period, according to the seasonality of mosquito-borne diseases. This study was conducted in the only health unit of the region, assessing patients reporting symptoms associated with acute viral infection by the dengue (DENV), Chikungunya (CHIKV) and Zika viruses (ZIKV), following laboratory con rmation by molecular and serological tests. In Brazil, studies involving LSS for arboviruses transmitted by Aedes sp. were carried out in large coastal cities where the population is concentrated, such as the Northeast [12,16] and Southeast regions [21,22], but not signi cantly replicated in the intra-continental portion [23,24].
The study area has previously been the object of epidemiological investigation based on cross-sectional interviews, reporting 28.6% of the symptoms associated with arbovirus infections in the sample population [25], albeit without laboratory con rmation. The present study aimed to determine the circulation of DENV, ZIKV and CHIKV in the study area; assess the clinical-epidemiological pro le of laboratory con rmations; estimate the probable infection site for laboratory con rmations, identifying possible autochthonous transmissions, and; given that the study area is geographically isolated and the health unit is the only facility treating the vulnerable population, compare LSS results with traditional epidemiological surveillance data. It is important to underscore that the study concluded in March 2020, before the expected 1-year observation period (June 2020), due to the COVID-19 pandemic.

Study Design
Event-based LSS de ned by acute arbovirus infection transmitted by Aedes sp. was conducted in a health unit, the only facility in the slum, and the rst destination of individuals seeking medical care. Based on the de nition of a probable case in the Manual for Adult and Child Diagnosis and Clinical Management for Dengue Fever [26], acute febrile and/or exanthematous illness was considered the de ned event for LSS, thereby opting for broad spectrum symptomatology that addressed the symptomatic infections of the three arboviruses investigated here: for DENV, probable case symptomatology is de ned as a sharp temperature increase (39-40ºC) accompanied by two or more of the following symptoms: headache, myalgia, arthralgia and retro-orbital pain, with exanthem present in 50% of symptomatic infections, as well as with anorexia, nausea, vomitting and diarrhea; for CHIKV, symptoms are similar, underscoring the intensity of arthralgia; for ZIKV, symptoms are characterized by absent or low to moderate-intensity fever, while exanthem occurs more frequently in the rst days of the infection.
The sample population consisted of residents in the study area who visited the health unit in the morning (from 8:00 am to 12:00 pm) between Monday and Friday, complaining of acute fever and/or exanthemathous illness (the de ned event). The primary identi cation of patients was conducted by family health teams (FHT) [27]. Each FHT is composed of a group of health professionals (doctor, nurse, community health agent) responsible for a geographic area within the respective region served by the health unit, which provides primary care. The health unit where LSS was carried out has 12 FHTs, which are instructed to refer patients presenting with the de ned event. After the patient was referred, the following inclusion criteria were established: being older than 18 years of age; being a resident of the study area (at least four days a week for at least three months before the visit to the health unit); and not exhibiting any contraindication for venipuncture. The exclusion criteria were being a homeless person and illiterate (given the need to provide written informed consent).
After the patients were referred by the FHT and had agreed to take part in the study, they signed the informed consent form and were given a brief explanation of the study objectives. Blood was collectedcharacterized as the "acute moment" -and 15-21 days after symptom onset, participants were contacted again for a second blood collection -characterized as the "convalescent moment" -provided they did not exhibit cough and/or coryza ( u-like upper respiratory tract syndrome) [28]. The laboratory methodology recommended by the Ministry of Health [26] was applied for molecular and serological con rmation: RT-PCR was used for the DENV (for serotypes), CHIKV and ZIKV in acute samples; and ELISA-IgM for DENV and CHIKV in both acute and convalescent samples. For the ELISA-IgM serological assay, two parameters were established: acute infection when seroconversion occurs, and recent infection when serology is positive in the acute sample [16]. To con rm acute infection, positive RT-PCR and/or seroconversion are considered laboratory evidence. ELISA-Igm was only tested for DENV due to the possible cross-reaction between dengue and Zika (both aviviruses).
Based on the seasonal pattern of mosquito-transmitted diseases, an initial 1-year observation period was established for the present study, between June 2019 and June 2020. However, given the lack of knowledge regarding SARS-CoV-2 infection at the time, as well as the lack of biosecurity for conducting research at the health unit, it was decided to halt the study in March 2020, after 9 months of observation.

Interview and Blood sampling
Blood collection at the acute moment was performed using a tube with separator gel (5 mL), followed by the interview to obtain demographic and socioeconomic characteristics (age, sex, ethnicity, address, domestic sanitary conditions); rst day of symptoms; clinical characteristics of the event; daily routine in the last 15 days before symptom onset (school, university, work or travel); the epidemiological link to an individual (family, neighborhood, work or school mate) with a positive medical diagnosis for DENV. Between 15 and 21 days after symptom onset, participants who did not report cough and/or coryza were contacted for convalescent blood collection at the conclusion of the second part of the questionnaire (focusing on symptoms reported at the acute moment: cure or persistence).
A standardized questionnaire from the REDCap website, version 7.5.0 (www.project-redcap.org/) was applied using mobile devices (tablets).

Laboratory Preparation
Tubes containing the collected blood were kept at ambient temperature in the health unit laboratory, followed by centrifugation of 1500 x g/10 min, aliquoting the supernatant serum in triplicate microtubes (usually 0.5 mL in each microtube). Once duly labeled with the respective patient code, the samples were initially kept in the laboratory freezer of the health unit (-20° C), then taken in thermal containers to the biorepository of the Center for Tropical Medicine, University of Brasilia (UnB) (-80° C), where they were stored until molecular and serological analyses.

Viral RNA extraction
Viral RNA extraction of 140 μL of serum from the acute sample was conducted using the QIAamp Mini kit, according to the manufacturer's recommendation (www.qiagen.com/us/resources), resulting in 60 μL of eluate.
Reverse transcription and arboviral RNA detection RT-PCR was performed using the "multiplex ZDC" kit from the Institute of Molecular Biology of Paraná with 38 μL of eluate, in accordance with the manufacturer's recommendation (www.ibmp.org.br/en-us/). In this protocol [30], a standardized 96-well microplate is subdivided into four isometric subgroups of 24 wells for the respective target arboviruses: DENV 1/4; DENV 2/3; CHIKV; and ZIKV. Each of these four subgroups on the plate is composed of 23 wells to test the samples (38 uL/4 subgroups= 9.5 uL of eluate for each subgroup), and 1 positive control from the kit itself. RT-PCR was ampli ed using the QuantStudio 5 real-time system (http://thermo sher.com/), and the results analyzed by the company's software.

Serological testing for anti-DENV IgM
Both acute and convalescent samples were tested for anti-DENV using the Panbio Dengue IgM Capture ELISA kit, according to the manufacturer's recommendation (www.globalpointofcare.abbott). Reading was done on a Kasuaki absorbance reader (450 nm), using Panbio units for cutoff values, following the criterion: <9 non-reactive; between 9 and 11 inconclusive; > 11 reactive for anti-DENV IgM.
Serological testing for anti-CHIKV IgM Both acute and convalescent samples were tested for anti-CHIKV using the Euroimun Chikungunya IgM ELISA kit, in line with the manufacturer's recommendation (www.euroimmun.com). Reading was done on a Kasuaki absorbance reader (450 nm), using relative units (RU/mL) for cutoff values, based on the following criterion: <0.8 RU/mL negative; between 0.8 and 1.1 RU/mL inconclusive; >1.1 RU/mL positive for anti-CHIKV IgM.

Data Analyses
The number of de ned events and laboratory con rmations was illustrated on epidemic curves, organized by epidemiological week (EW) starting with symptom onset. Clinical-epidemiological characteristics were cross-referenced with laboratory con rmations, using frequency tables. Information on routines for the 15 days prior to symptom onset was used to estimate the probable infection site.
In order to compare LSS results with traditional epidemiological surveillance data, probable arbovirus cases noti ed by the Cidade Estrutural and the Federal District were extracted from the Ministry of Health (datasus.gov.br) and Federal District Health Department websites (info.saude.df.gov.br): for the former, the interface allows extraction per period (year, month, EW) and nal classi cation of the probable case (con rmed, discarded, not investigated), but not separation by the ARs of the Federal District; for the latter, the interface makes it possble to extract by period and site (including AR), but only dengue data are made available. Epidemiological bulletins from the Federal District Health Department were consulted to complement this information [31]. The nal classi cation criteria of the probable case (clinicalepidemiological or laboratory) for Cidade Estrutural was obtained; the laboratory methods used (rapid NS1 test; IgM serology; PCR; viral isolation); and the dates of the rst symptoms and blood collections for the respective laboratory tests established. Epidemic curves of probable dengue and Chikunguya cases from traditional surveillance for Cidade Estrutural were organized by EW starting on the day of symptom onset for the study period, discriminating the respective nal classi cation (dengue, Chikunguya, discarded and non-investigated). Probable Zika cases and the respective nal classi cation were not arranged by EW for the study area due to the non-availability of these noti cations at the AR level for the Federal District. In order to observe the statistical differences between traditional surveillance and LSS results, the Wilcoxon test for paired data was conducted. To contextualize the COVID-19 situation in the rst quarter of 2020, Federal District Health Department epidemiological bulletins from the Center of Emergency Operations for Covid-19, established in February of that year, were consulted [32].
Microsoft Excel was used to manage the datasets and create tables and graphs, Stata 14.0 software for statistical analyses, and Google Earth for satellite images. The GPS points of the dwellings were collected using a Garmin Etrex 10 handheld navigator.

Results
The study period was between the third week of June (EW 25) 2019 and third week of March (EW 12) 2020. A total of 157 individuals who presented with the de ned event were referred by the FHT, 134 of whom were considered eligible for the study. The information on 3 of these was not saved electronically. At the end, 131 de ned events were identi ed by the study. For the second blood collection (convalescent moment), the samples of 36 individuals were collected, given that 73 de ned events exhibited cough and/or coryza; 15 individuals refused to participate, and 9 were not approached due to the early termination of the study.
Demographic and socioenvironmental characteristics are presented in Table 1. The LSS population samples consisted of adults aged around 30 years, with approximately 10 years of schooling, most of whom were non-white (89%) and women (62.6%). Most of the population sample did not have private health insurance (94.6%) and approximately 4 of 5 individuals reported that their house has environmental sanitation (80%), an indoor plumbing system (79%) and trash collection (81%). However, 40% declared having to ration water and 34.3% stored water by other methods in the three months before the interview.    The clinical-epidemiological characteristics of the 132 de ned events and respective laboratory con rmations are shown in Table 2. Nearly all of the ve most common symptoms associated with acute viral infections (headache, fever, myalgia, arthralgia and retro-orbital pain) were observed in positive laboratory con rmations, except for one DENV-1 without arthralgia and two DENV-2 without arthralgia and retro-orbial pain, respectively. Cough and coryza were observed in one DENV-1 and one DENV-2 case, respectively. Exanthem was present in DENV-IgM (identi ed in June 2019) and two DENV-1 con rmations (identi ed in January-March 2020). CHIKV stands out for the presence of sore throat. Although alarming signs were reported (di culty breathing, bleeding), no severe dengue fever was found. The main characteristics related to the daily routines and possible exposure for laboratory con rmations by LSS are presented in Supplementary Material 2. Individuals from samples 39 (DENV-2), 98 (CHIKV) and 106 (DENV-1) were not exposed outside their homes, and it is plausible to characterize these cases as household transmission. For individual 27 (DENV-IgM), Cidade Estrutural can also be considered the probable infection site (residence or workplace). The workplace of individual 37 (DENV-2) is in the neighboring AR (SCIA, see Figure 2-C), but since this individul works only at night, when Aedes aegypti activity declines signi cantly [33], household transmission is also suggested in this instance.   With respect to statistical analyses between traditional surveillance data and LSS results, considering a 95% CI and the null hypothesis that the magnitudes of SE-paired values are equal, differences were observed for all the comparisons performed: between probable dengue cases and de ned events (p=0.0346); between dengue cases and de ned events (p=0.0238); and between dengue cases and laboratory con rmations (p=0.000). In terms of the traditional surveillance criteria used to con rm The study found evidence of two DENV serotypes and CHIKV circulating, with high plausibility of at least three individuals having been infected in their homes (positive for DENV-1, DENV-2 and CHIKV). Cidade Estrutural is hyperendemic for arboviruses transmitted by Aedes sp. mosquitos, with the population exposed to the risks of secondary infection, co-infection and clinical outcomes in areas where this "arbovirus soup" occurs [15], despite cross-immunity being observed among avivirus infections [40;41].
The site is a potential area for the spreading of these arboviruses to other ARs, given the cosmopolitan presence of Aedes aegypti throughout the Federal District and the human mobility that occurs in the region.
The possible low number of laboratory con rmations by LSS for the population sample may be due to the inclusion of nonspeci c symptoms in the clinical spectrum of the de ned event, including upper respiratory tract symptoms (cough 66; coryza: 62; both 53) which are not associated with the de nition of probable cases for arboviruses transmitted by Aedes sp.
[26]. This option was adopted to maximize sampling for detection by laboratory methodologies, and to observe the clinical pro le of laboratory con rmations. However, when compared to other LSS studies on arboviruses transmitted by Aedes sp. conducted in health units in the country, the number of positive DENV cases observed here (6.1%) was higher than that reported by Silva [24]. For the other two arboviruses, the other studies obtained higher results. In addition to the particularities of each investigation and the study population, this discrepancy depends on whether outbreaks and epidemics occur while LSS is being carried out, as occurred in the study by Silva et al. 2019 [16] during the 2015 Zika epidemic in Salvador, Bahia state.
Laboratory con rmations exhibited the ve most common symptoms involving acute viral infections, con rming the low speci city for diagnosing arbovirus infection when not supported by laboratory analysis [16]. Cough and coryza were observed in 1 DENV-1 and 1 DENV-2 con rmation, respectively. This aspect of acute viral syndromes is used to discriminate airborne viruses from others (such as arboviruses) [28]. However, albeit uncommon, upper respiratory tract symptoms may be associated with acute arboviral infections transmitted by Aedes sp., including cough, coryza and nasal congestion [42]. It is important to underscore that DENV has been isolated in secretion samples collected from the upper respiratory tract, raising the hypotheis of possible transmission via aerossol or close contact, as observed in the replication of this virus in culture mediums of airway epithelium cells. However, evidence is still insu cient to con rm this infection pathway [43].
Given the acute DENV infection con rmed only for seroconversion (DENV-IgM), the possible crossreaction between aviviruses in immunoenzymatic assays for IgM, and that exanthem is associated with acute ZIKV infection, suggest the transmission of this virus. The low viral load observed for ZIKV in blood serum, when compared to DENV, [44] reinforces the suggestion of infection by the former (which also explains the more intense fever in DENV than in ZIKV infection). For ZIKV, testing in other biological specimens such as urine was recommended (especially for pregnant women suspected of being infected) [45]. Although this seroconversion has been identi ed in the DENV-2 cluster (between EW 25 and 29 of 2019), suggesting infection of this dengue virus serotype, none of these DENV-2 exhibited enxanthem, only con rmations for DENV-1 located in the other cluster (between EW 4 and EW 12 of 2020 The obvious contrast between dengue cases noti ed by traditional surveillance and LSS con rmations indicates that different acute viral infections may be noti ed as dengue. The absence of con rmed noti cation of Chikungunya, with probable cases of this virus classi ed as dengue, suggests underreporting. The con rmation criteria of probable cases reported by traditional surveillance for the study area and the laboratory methodologies used demonstrate the inconsistencies for 2019 [46], namely that the laboratory criteria value (355) was lower than that of rapid NS1 tests (360). With respect to laboratory criteria sensitivity, it is known that the rapid tests routinely used for point-of-care in outpatient facilities may exhibit low sensitivity when compared to those obtained in the laboratory [47]. Given the limitations of the clinical-epidemiological criterion, and dependence on an effective ow for laboratory con rmation of infectious diseases, it can be suggested that different acute infections such as DENV are noti ed for the study area, while other arboviruses are underreported. Several areas of Brazil have a vulnerable population that are dependent on the National Health System (SUS, in Portuguese), which has undergone a structural and nancial decline in recent years [48]. The worst scenario has been observed during the COVID-19 pandemic, considering the areas with active arbovirus transmission [49].
The subsequent endemic circulation of SARS-CoV-2 in the Brazilian population will be a further hindrance to the noti cation of acute viral syndromes when not con rmed by laboratory criteria.
A number of limitations were observed in the Cidade Estrutural health unit. First, communication with different FHTs produced heterogeneous identi cation of de ned events during the study period, due to limitations of the teams themselves. Second, the LSS could not be maintained during the daily operation of the health unit (8:00 am-6:00pm), and not collecting the convalescent sample of some patients contributed to reducing the sampling effort for laboratory analysis. Third, the COVID-19 pandemic precluded concluding the 1-year observation period and delayed the laboratory processing of biological samples, since resources and personnel were redirected to combat the new virus. The pandemic can also be considered a potential bias for the FHT in identifying de ned events and traditional surveillance data in 2020. However, Supplementary Material 3 and the epidemiological bulletin for COVID-19 [32] reported the rst con rmed noti cation on March 5 for an individual who had traveled to Europe; the second con rmation occurred on March 10 (wife of the rst case); 19 con rmations on March 16; and 87 on March 20, 5 of which classi ed as local transmission, meaning that January and the rst three weeks of February (SE1 to SE7) can still be considered pre-pandemic period. This is also supported by the Ministry of Health's recognizing the existence of community transmission of SARS-CoV-2 on March 20 [32]. In this pre-pandemic context, LSS identi ed DENV-2 (SE 4) and CHIKV (SE 7).
LSS results raise questions regarding the time elapsed between laboratory con rmations of DENV-2 and DENV-1 during the second half of 2019, and the absence of con rmations between these two clusters.
Are these transmissions sustained at the local level in the study area during low Aedes sp. density, depending on the seasonality of Aedes aegypti-borne diseases? In addition, does any type of ecological substitution of DENV serotypes occur during this seasonality? Thus, to answer these questions, active surveillance should be performed (also focusing on asymptomatic individuals), along with entomological surveillance and analyses of mosquito infectivity.

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.
Competing interests methodology; PRM, LAGN, TSCQ, TFN, DCCA and RH performed laboratorial analysis; PRM interpreted the results, wrote the original draft of manuscript and designed the visualization; PRM, TSCQ, TFN, RH, GASR and WNA reviewed and edited the nal manuscript. All authors read and approved the nal manuscript.