The burden of dengue fever and chikungunya in southern coastal Ecuador: Epidemiology, clinical presentation, and phylogenetics from a prospective study in Machala in 2014 and 2015

Here we report the methods and findings from an arbovirus surveillance study conducted in the city of Machala, Ecuador, from January 1, 2014 to December 31, 2015. Patients with suspected DENV infections (index cases, n=324) were referred from five Ministry of Health sentinel clinics. A subset of DENV positive index cases (n = 44) were selected, and individuals from the index household and four neighboring households within a 200-meter radius (associates) were recruited (n = 400). In 2014, 70.9% of index cases and 35.6% of associates had evidence of acute or recent DENV infections. In 2015, 28.3% of index cases and 12.8% of associates had acute or recent DENV infections. For every DENV infection detected by passive surveillance, we detected an additional three acute or recent DENV infections in associates. Of associates with acute DENV infections, 68% reported dengue-like symptoms, with the highest prevalence of febrile acute infections in children under 10 years of age. The first CHIKV infections were detected on epidemiological week 12 in 2015. 41% of index cases and 4.6% of associates had acute CHIKV infections. No ZIKV infections were detected. Phylogenetic analyses of isolates of DENV from 2014 revealed genetic relatedness and shared ancestry of DENV1, DENV2 and DENV4 genomes from Ecuador with those from Venezuela and Colombia, indicating presence of viral flow between Ecuador and surrounding countries. Enhanced surveillance studies, such as this, provide high-resolution data on the distribution of symptomatic and subclinical arboviral infections across the population.

In Ecuador, suspected and confirmed DENV, ZIKV, and CHIKV cases require indoor residual spraying, source reduction, and larvicide application). There have been prior enhanced surveillance studies to estimate the burden of dengue population. This is especially important in settings and in subgroups with low-health care 1 1 3 seeking behavior or limited access to health centers. These data allow the public health sector to 1 1 4 more accurately estimate the social and economic burden of the disease, allowing for more vaccination campaigns or vaccine trials. were informed that they might be randomly selected to participate in a cluster investigation. Sequencing and consensus assembly.

7 9
Samples from 2014 that were DENV positive by RT-PCR were sent to Walter Reed as input DNA, neutralization of each Nextera® XT Tagmentation reaction using 5µl NT buffer, with Illumina protocols. fastq data were stripped of barcodes and adapters and subjected to read filtering using a quality threshold of Q25. Remaining reads were further end-trimmed using a quality threshold of Q25 were constructed, in parallel, from the adapted VCF output. All consensus sequences were The five sequenced full genome DENV1 samples were aligned to a set of full genome Likewise, a set of 100 full genome DENV4 reference sequences was obtained from GenBank 3 2 8 following the same criteria as for DENV1, and aligned to the single DENV4 sequenced genome 3 2 9 from Ecuador. We were unable to sequence DENV3 due to limited sample volume. Genetic  The best-fit models of evolution for DENV1, DENV2 and DENV4 datasets were and Regrafting). Node confidence values were determined by aLRT (approximate Likelihood this study (Table 1). A total of 72 index cases were positive by NS1 rapid test, and from these we results and were included in this study.  detected. CHIKV was first identified in our study on epidemiological week 12 in 2015, and 3 5 7 transmission followed a similar seasonal curve as DENV (Fig 3). No ZIKV infections were 3 5 8 detected (Table 1).

7 3
In 2015, more index cases were positive for acute CHIKV (50/122, 41%) than for acute  We estimated the prevalence of febrile acute (FA) infections for DENV and CHIKV by   (Table 5). Index cases with acute DENV infections were significantly more likely to report 3 9 5 anorexia and nausea, vomiting, and abdominal pain, whereas index cases with CHIKV were 3 9 6 more likely to report rash (p<0.05).

9 7
We also compared the demographics and symptoms of primary versus secondary DENV 3 9 8 infections, and DENV1 versus DENV2 infections in index cases. Overall, we identified more 3 9 9 severe illness in secondary DENV infections than in primary infections (Supplementary Table   4 0 0 4). Individuals with secondary DENV infections were significantly older, were more likely to be 4 0 1 hospitalized, and were more likely to report vomiting (p<0.05). Individuals with primary DENV 4 0 2 infections were more likely to report fever (p<0.05). We did not find significant differences in  Table 5). In each cluster of homes, approximately nine associates were recruited into this study per   infection (Table 4). There were no significant differences in the demographics or febrile  For associates, we determined the prevalence of FA infections for DENV by age class  We report the expansion factors (EF) for acute and recent DENV infections, acute DENV 4 3 8 infections, and symptomatic (i.e., one or more dengue-like symptoms) acute DENV infections 4 3 9 ( Table 6). The EF calculations were adjusted to account for two associates with acute DENV The best-fit models for the evolution of DENV1, DENV2, and DENV4, as determined by of DENV1 genotypes I, II, IV and V, and the sylvatic genotypes III and VI (Fig 4). The five  The ML phylogenetic tree of DENV2 showed a clear distinction of DENV2 genotypes, including sylvatic, American, Cosmopolitan, Asian I, Asian II and Asian/American (Fig 5). The  IIA, IIB, III and sylvatic (Fig 6). However, two taxa from India/1961-1962 clustered with for the Ecuador node was low (0.37) suggesting that its correct placement was uncertain. In this study, we characterized the epidemiology and clinical characteristics of CHIKV   The expansion factor (EF) for DENV in Machala was estimated using acute/recent end of a range of previously reported EFs for the PAHO region. 39 Interestingly, we found that the  times more dengue infections than were reported to the national surveillance system. 53 A study in 5 1 2 Peru compared passive surveillance of dengue to a cohort study and sentinel clinic surveillance, They found that both sentinel and cohort surveillance methods detected an increase in dengue One of the limitations of this study was that we surveyed the nearest neighbors of the randomly within the 200-meter radius for a more accurate measure of disease prevalence.

2 3
Additionally, this study was limited to five clinical sites operated by the MoH that were willing and able to support the study. Burden of CHIKV and other febrile illness: In 2015, we found that 41% (50/122) of clinically diagnosed DENV infections (index 5 2 8 cases) were positive for CHIKV, higher than the proportion of laboratory-confirmed DENV 5 2 9 infections. We identified six index cases (6/122=4.94%) and one associate (1/87=1.1%) with burden of CHIKV is likely higher than reported here, since anti-body tests were not utilized. This other febrile illnesses are co-circulating. These data also suggest that the large increase in DENV other circulating arboviruses, including CHIKV.

3 8
We did not detect ZIKV in our surveillance system during the study period, consistent more recent studies shown that ZIKV may be more readily detected in urine and whole blood, Clinical predictors of DENV and CHIKV.

4 7
In general, the symptoms that were observed with acute DENV infections in this study burden of CHIKV in Machala was among adults aged 20 to 49. We found that rash was more commonly reported by people with CHIKV infections than those with DENV, which is We found that two thirds of associates with acute DENV infections were symptomatic; reported that men in the urban periphery were less likely to seek healthcare, 18 and medical subclinical infections and infections in demographic groups who do not seek health care. Phylogenetic analyses of DENV1 showed Ecuadorian samples falling into two distinct  from Ecuador were most closely related to genomes from Venezuela and Colombia. However, between DENV4 in Ecuador, Venezuela, Colombia and Brazil. It is important to note that 5 9 0 samples from Peru were missing here as well, and that there is a possibility this country was also DENV between Ecuador and surrounding countries, including introduction and re-introduction DENV movement and spread in this region. This study contributes to a long-term collaboration with the MoH and other governmental  Enhanced surveillance studies, such as this, provide high-resolution spatiotemporal data  This is especially important in places and in subgroups with low healthcare seeking behavior, interventions, such as vector control, community mobilization, and vaccines. The age-stratified 6 1 5 prevalence data generated through this study design also provides important information for the 6 1 6 design of vaccine trials and vaccination campaigns. Acknowledgements. This project was possible thanks to support from colleagues from the 6 1 8 Ministry of Health, the National Institute of Meteorology and Hydrology, the National Secretary 6 1 9 of Higher Education, Science, Technology, and Innovation (SENESCYT) of Ecuador and 6 2 0 community members from Machala, Ecuador. We thank our local field team and coordinators for analysis, and data compilation. We also thank Dr. Renato Leon for supporting the development  Disclosures. The authors declare no competing interests, financial or non-financial.     FA infections divided by all individuals from the age class who were recruited into the study and  recruitment, or self-reported fever within the last 7 days. See Supplemental Table 5 for raw data.