https://orcid.org/0000-0002-5103-2278
Figures
Abstract
Background
Aedes aegypti presence, human-vector contact rates, and Aedes-borne virus transmission are highly variable through time and space. The Lower Rio Grande Valley (LRGV), Texas, is one of the few regions in the U.S. where local transmission of Aedes-borne viruses occurs, presenting an opportunity to evaluate social, urbanistic, entomological, and mobility-based factors that modulate human exposure to Ae. aegypti.
Methodology & Principal findings
Mosquitoes were collected using BG-Sentinel 2 traps during November 2021 as part of an intervention trial, with knowledge, attitudes, and practices (KAP) and housing quality surveys to gather environmental and demographic data. Human blood samples were taken from individuals and a Bitemark Assay (ELISA) was conducted to quantify human antibodies to the Ae. aegypti Nterm-34kDa salivary peptide as a measure of human exposure to bites. In total, 64 houses were surveyed with 142 blood samples collected. More than 80% of participants had knowledge of mosquito-borne diseases and believed mosquitoes to be a health risk in their community. Our best fit generalized linear mixed effects model found four fixed effects contributed significantly to explaining the variation in exposure to Ae. aegypti bites: higher annual household income, younger age, larger lot area, and higher female Ae. aegypti abundance per trap night averaged over 5 weeks prior to human blood sampling.
Conclusions
Most surveyed residents recognized mosquitoes and the threat they pose to individual and public health. Urbanistic (i.e., lot size), social (i.e., income within a low-income community and age), and entomological (i.e., adult female Ae. aegypti abundance) factors modulate the risk of human exposure to Ae. aegypti bites. The use of serological biomarker assays, such as the Bitemark Assay, are valuable tools for surveillance and risk assessment of mosquito-borne disease, especially in areas like the LRGV where the transmission of target pathogens is low or intermittent.
Author summary
Aedes aegypti is a mosquito vector with public health importance on the global scale as it transmits viruses such as dengue, chikungunya, and Zika. Although transmission rates of dengue and Zika are low in the U.S., there are a few regions, including south Texas, where local transmission has occurred. Our study aimed to evaluate the factors associated with risk of exposure to these viruses using a serological bioassay that measured antibody response to an Ae. aegypti salivary protein to assess human-vector contact. We collected mosquitoes, took human-blood samples, and conducted urbanistic and demographic surveys in November 2021 in eight communities in the Lower Rio Grande Valley, Texas. Our knowledge, attitude, and practices survey found that most residents recognized adult mosquitoes, though few individuals knew someone personally who been sick with a mosquito-borne disease. Outdoor adult female Ae. aegypti abundance was positively associated with exposure to mosquito bites. Household income, individual age, and lot area also significantly affected exposure levels. The Bitemark Assay we used in this study can be utilized as a tool for entomological risk assessment and could be used as an alternative to infection exposure in areas where mosquito-borne disease levels are low.
Citation: Scavo NA, Juarez JG, Chaves LF, Fernández-Santos NA, Carbajal E, Perkin J, et al. (2024) Little disease but lots of bites: social, urbanistic, and entomological risk factors of human exposure to Aedes aegypti in South Texas, U.S. PLoS Negl Trop Dis 18(10): e0011953. https://doi.org/10.1371/journal.pntd.0011953
Editor: Philip J. McCall, Liverpool School of Tropical Medicine, UNITED KINGDOM OF GREAT BRITAIN AND NORTHERN IRELAND
Received: January 30, 2024; Accepted: September 29, 2024; Published: October 21, 2024
Copyright: © 2024 Scavo et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are in the manuscript and its supporting information files.
Funding: This research was funded by the Centers for Disease Control and Prevention, contract 200-2017-93141 and Texas A&M AgriLife Research that were both received by GLH and supported NAS, JGJ, NAF, EC, and GLH. NAS was also supported by a Texas A&M University, Association of Former Students Graduate Merit Fellowship. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the Centers for Disease Control and Prevention or the Department of Health and Human Services. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
The yellow fever mosquito, Aedes aegypti L. (Diptera: Culicidae), is the main vector of arboviruses such as dengue, Zika, and chikungunya viruses. Diseases caused by these Aedes-borne viruses pose a threat to global health with dengue virus (DENV) affecting 390 million people annually [1], Zika having autochthonous transmission in 87 countries [2], and chikungunya causing an average annual loss 106,000 disability-adjusted life years (DALY) [3]. With a lack of safe vaccines that are effective against primary infections, public health authorities rely on controlling the mosquito vectors to reduce disease. Aedes aegypti vector control has shown variable levels of efficacy in reducing mosquito populations [4,5] and preventing disease transmission of arboviruses [6–8]. With the risk of a shifting global distribution of Ae. aegypti mosquitoes [9] and some regions showing potential for vector presence without associated viruses [10], novel methods of surveillance and control that focus on potential areas of arboviral emergence areas are needed.
The region known as the Lower Rio Grande Valley (LRGV) along the U.S.-Mexico border in south Texas is one of the few regions in the U.S. with local vector-borne transmission of Aedes-borne viruses [11], and more recently human malaria [12]. This can be explained by the fact that Ae. aegypti is well-established in the area as an efficient urban vector due to its affinity for man-made container habitat and highly anthropophilic behavior [13]. Colonias are low-income, mostly Hispanic communities in unincorporated areas leading to a general lack of services (e.g., poor water sanitation) [14] and marginalized residents with little political power [15]. Understanding the ecological and social factors that modulate Ae. aegypti abundance and human exposure to mosquito bites is key to developing efficient and effective vector control programs. Some of the previously detected risk factors for an increased indoor and outdoor abundance in low and middle-income communities of the LRGV include demographic indices (i.e., number of children and toddlers) and housing variables (i.e., air conditioning window-mounted units, number of windows) [16]. Furthermore, the presence of air-conditioning reduced the risk of prior exposure to dengue virus in this region [17]. More generally, household water management strategies [18], socio-economic factors (SEF) [19–22], and climate variables [23,24] have been tied to presence or abundance of Aedes mosquitoes. However, we currently lack the understanding of how some of these risk factors relate to human-vector interactions with the added component of human mobility or how vector abundance translates to exposure to vector bites.
Current guidelines for evaluating the success of a vector control intervention recommend the use of an epidemiological endpoint, such as active infection or past exposure to pathogens, to inform intervention efficacy [25]. However, many regions with local transmission of mosquito-borne pathogens do not have consistent and sufficient burden of human disease necessary for an epidemiological endpoint. For instance, a West Nile virus (WNV) serosurvey in humans in Connecticut, U.S. didn’t find seropositive participants in their study despite WNV being prevalent in neighboring states and that birds and mosquitoes were infected in the region [26,27]. Similarly, the LRGV represents the margin of Aedes-borne virus endemicity and has only sporadic local transmission of DENV (24 locally acquired cases from 2010–2017) [28]. The same pattern exists for malaria transmission, with a single autochthonous human case of malaria in South Texas in 2023 [12]. This makes evaluating an intervention using an epidemiological endpoint variable difficult to utilize in a priori planning of a vector control intervention study.
To overcome this limitation, the use of a human antibody response to mosquito salivary proteins has emerged as a valuable tool [29,30] to be used as a complementary endpoint which combines elements of an entomological and epidemiological endpoint. In this context, humans develop antibodies in response to exposure to salivary proteins associated with vector bites, and immunological assays can detect this past evidence of exposure to vector bites. Progress measuring human antibody response to vector bites has advanced from using crude salivary gland proteins (e.g., whole saliva) [31,32] to more recent progress identifying peptides specific to different mosquito taxa, improving assay specificity [33,34]. Moreover, such tools can be used to estimate the risk of arbovirus exposure in areas with high transmission [33,35] and are a more direct measure of risk than adult female mosquito abundance. Although adult female abundance has been correlated with dengue incidence in some cases [36,37], there are some limitations to this approach. For instance, human landing catches cannot be ethically performed with Aedes species [38] as there are no universally effective vaccines for Aedes-borne viruses, except for yellow fever virus. Because of this, mosquito sampling of host-seeking, oviposition seeking, or resting populations are used but these do not measure mosquito contact with humans and thus do not account for mosquito or host behaviors (e.g., use of spatial or personal repellents), and which have different levels of efficiency [39].
It has been shown that for different mosquito species the use of IgG antibody response to mosquito salivary gland proteins can serve as effective indicators of human-vector contact as an exposure biomarker [34,40–42]. Moreover, high bite exposure measured via this method has been linked to disease levels in humans for both the malaria [34,40] and dengue systems [33,43] indicating its usefulness in assessing mosquito-borne disease risk. IgG antibodies are used over IgM as their specificity is greater [33], and is specific to the genus level [44], with low levels of cross-reactivity between Ae. aegypti and Ae. albopictus [45].
Even so, immune response intensity is not always linked to the likelihood of being bitten by infected mosquitoes [38]. Although a higher immune response likely means more bite exposure, it might not directly translate to be being bitten by a mosquito that is positive for dengue. In Fustec et al. [38], the direct relationship of human dengue infections and antibody response to Aedes salivary proteins could not be measured as no dengue cases were detected during the study period [38], and other studies have shown a link between exposure levels and dengue cases in humans [33,43]. Salivary biomarkers also allow for individual mosquito bite exposure assessment and improve the ability to assess heterogeneity of disease transmission compared to community-level entomological measures [42].
Our study focuses on comparing the use of an antibody response against Ae. aegypti salivary gland peptides (i.e., Nterm34kDa) as an endpoint measurement in relation to mosquito abundance. Also, we aim to describe social, urbanistic, and human mobility risk factors associated with Ae. aegypti abundance and exposure to their bites in low-income communities (a.k.a. colonias) of the LRGV. The results build on our previous work in the area to elucidate seasonal patterns of mosquito abundance [46], dispersal of Ae. aegypti from discarded containers [47], and evaluating vector control interventions [5]. Ultimately, our work can help guide public health programs to better understand the local ecology of mosquitoes along the U.S.-Mexico border and how vector control interventions might be evaluated in the LRGV and in other regions.
Methods
Ethics statement
This project received approval from the Institutional Review Board of Texas A&M University (IRB2021-0886D). We obtained individual written consent from each household owner for the weekly outdoor entomological surveillance and KAP surveys. We obtained individual written consent from adults that participated in the blood sampling and assent from children, as well as written consent from a guardian, for the same procedure.
Study location and site selection
The study was carried out in the county of Hidalgo, Texas, U.S., which is part of the LRGV region located along the U.S.-Mexico border. The county of Hidalgo has an estimated 870,000 inhabitants, of which 92% consider themselves Hispanic or Latino origin, 26% are foreign borne individuals and 24% live in poverty (based on income and family size/composition) [48]. The climate in this region is considered humid sub-tropical, with a cold/dry season from November to February (7–21°C), and a rainy season that starts in April (18–30°C), peaks in September (23–33°C) and finishes in October (19–31°C) [49].
Sites were selected based on previous work in the area [16,46] where rapport had been built with community members. Briefly, potential sites were selected based on average income level per household, total number of households in the community, isolation of community, and distance from our base of operations in Weslaco, Texas. We selected eight low-income communities based on high community participation in past studies (Fig 1).
1 –La Piñata, 2 –South Donna, 3 –Balli, 4 –Progresso, 5 –Chapa, 6 –Mesquite, 7 –Indian Hills West, 8 –Indian Hills East. County and state boundaries from US Census Bureau’s Cartographic Boundary Files (https://www.census.gov/geographies/mapping-files/time-series/geo/carto-boundary-file.html) and the base map is the OpenStreetMap which is available under the Open Database License (https://tile.openstreetmap.org/{z}/{x}/{y}.png).
Entomological sampling
Adult mosquito sampling was carried out as part of an Auto-Dissemination Station (ADS) (BanfieldBio Inc.) intervention study, similar to a previous trail using a different ADS prototype [50]. The intervention study was a cluster randomized control trial that was carried out from June 2021 until March 2022. Surveillance was not conducted for an entire year (52 weeks) given that the risk of Aedes-borne virus transmission such as DENV and ZIKV in the LRGV peaks from September to November when the Ae. ageypti is most abundant, this is common practice in the area [51,52]. Mosquito sampling was done using BG Sentinel 2 traps (Biogents, Germany) baited with BG lures (Biogents, Germany) placed in the peridomicile (i.e., outside) of homes at a density of 1 trap per 500m2, with a up to 84 traps set per week, over 39 weeks. A total of 89 household participated at some point in the mosquito sampling. Traps were left for 24 hours once per week, collected mosquitoes were separated by sex and species (Ae. aegypti, Ae. albopictus, Culex sp., Anopheles, and other mosquitoes) and stored at -20°C. Mosquito identification was done based on morphology using taxonomic keys [53]. Households within communities were randomly selected based on desired trap density. Houses were approached for participation in the study and if homeowners agreed, a trap was placed in their lot. If a household dropped out of the study, a neighbor was recruited as a replacement in the following order: neighbor to the right, neighbor to the left, neighbor in behind, neighbor directly across the street.
Surveys and blood sample collection
Knowledge, Attitudes, and Practice (KAP) and housing quality surveys were conducted concurrently at the same household visit from November 1–13, 2021. The format of the surveys was similar to those previously done by our group in 2017 and 2018 [16]. Briefly, we used a structured face-to-face questionnaire with a mixture of close-ended, semi-closed-ended, open-ended and ranking questions. The KAP survey asked participants about household demographics, mosquitoes, the pathogens mosquitoes transmit, and their household member movement patterns. The house quality survey consisted of evaluating the quality of windows, doors, and their screens; housing construction material; and water holding containers and was carried out counterclockwise from the main house entrance. We recorded housing materials (timber/metal, cement, brick), screen quality (with holes, with no holes, size of holes) on windows and doors, the type of air conditioning (A/C) unit (window mounted, central) if present, and the type and number of mosquito container habitats found in each household peridomicile (see S1 Data). We defined the peridomicile of a household as the area between the property limit to the main house perimeter. From 95 houses under weekly entomological surveillance, 47 agreed to participate with the KAP, blood sample and housing surveys. An additional 17 houses were recruited as replacements (replacement houses were recruited from the left, right, back and front, adjacent to the BG house, if no house agreed no further house was recruited) for a total of 64 houses surveyed.
Blood sampling was done via finger prick and four circles of blood were collected on a 903-protien saver card (GE Healthcare, USA). Samples were dried and placed into a plastic bag with desiccant and stored at 4°C until further processing.
Bitemark assay
The An. gambiae gSG6-P1 peptide for the measure of exposure to Anopheles spp. [34] and the Ae. aegypti peptide Nterm-34kDa [54] were synthesized by Genscript (Piscataway, NJ, USA). ELISA conditions were standardized as described elsewhere [34,43]. Briefly, dried blood samples were prepared by punching a 6mm circle out of the Whatman 903 protein saver card (GE Healthcare, US), and eluting it into 500 μL of elution buffer (PBS 1×) and incubating overnight at 4°C. At the time of sample preparation UltraCruz High Binding ELISA Multiwell Microplates (96-well) were coated with 100 μL/well of either gSG6-P1 or Nterm-34kDa peptide (2 μg/mL). Plates were incubated overnight at 4°C and blocked with 200 μL of 5% skim milk solution in PBS-tween 20 (0.05%) (Blocking buffer) for 30 min at 37°C. The sample elution was used to prepare a 1:50 dilution of the sample in blocking buffer. Then, 100 μL of that dilution were added to each well (individual samples were tested in duplicate). Plates were incubated at 37°C for 2 h, washed three times, then incubated 1 h at 37°C with 100 μL/well of a 1/1000 dilution of goat monoclonal anti-human IgG conjugated with horseradish peroxidase (ABCAM, Cambridge, MA). After three final washes, colorimetric development was carried out using tetra-methyl-benzidine (Abcam) as a substrate. In parallel, each assessed microplate contained in duplicate: a positive control (pool of diluted samples), a negative control (wells with no human sample), and a blank (Wells with no antigen). The blank was composed by wells containing no sample. The reaction was stopped with 0.25 N sulfuric acid, and the optical density (OD) was measured at 450 nm.
Optical density normalization and plate to plate variation was performed as previously described by our group and others [34]. Briefly, antibody levels were expressed as the ΔOD value: ΔOD = ODx − ODb, where ODx represents the mean of individual OD in both antigen wells and ODb the mean of the blank wells. For each tested peptide, positive controls of each plate were averaged and divided by the average of the ODx of the positive control for each plate to obtain a normalization factor for each plate as previously described [34]. Each plate normalization factor was multiplied by plate sample ΔOD to obtain normalized ΔOD that were used in statistical analyses.
Statistical analysis
We collected 99 variables between the KAP and housing surveys. To make the analysis more manageable, we used dimension reduction methods to generate three indices (windows, doors, and hosts) following the procedures described in Chaves et al. [55]. To start, we carried out descriptive statistics on the KAP and housing datasets to assess which variables had low standard deviation which would impact the results of our data reduction techniques. We removed variables that fell under these categories, that were colinear in nature, or that had a high degree of missing values not collected in the original surveys. The door and window indices were generated via Principal Components Analysis (PCA) by grouping variables according to their relevance to door, window, and host categories. Other indices were tested, though only these three were kept as they explained more than 50% of the cumulative variability. Details on index creation and PCAs biplots can be found in S1 Statistical Analysis.
After data reduction and elimination of identification variables (e.g., street name, latitude, longitude) from the dataset, 47 explanatory variables remained. We then chose to select 10 variables from this set so that our n/k value, where n is the number of data points and k is number of parameters, would be above ten [56]. We used our knowledge of the literature to identify 10 variables from our set that were relevant to our study [56,57]. Rows with missing data were omitted from the analysis.
Our outcome variable of interest was individual exposure to Ae. aegypti bites measured via the Bitemark Assay (i.e., ΔOD). We analyzed how social, urbanistic, entomological, and movement factors were associated with bite exposure using a generalized linear mixed model (GLMM) approach. A mixed model was chosen to account for the potential lack of spatial independence (i.e., individuals nested within households and households nested within communities). While both linear mixed models [38,58] and GLMMs [59–61] have been used in similar serological-biomarker studies, we chose a Gaussian GLMM because they are an ideal tool for analyzing normal data whose independence is constrained by different factors (e.g., spatial or temporal) and that are modeled as random effects [62,63] in our study.
We constructed a global model (mglobal1B) to evaluate the effect of 10 fixed effects on ΔOD while controlling for non-independence among houses surveyed in the same communities. The 10 fixed effected included average distance in miles traveled per week, income (2 levels: <$25,000, >25,000), host community index (host.1), door index (door.2), age (years), sex (2 levels: male, female), AC type (4 levels, window, central, mini-split, none), average abundance of Ae. aegypti females averaged over 5 weeks prior to sampling, area of lot (m2), and total containers in the lot (Table A in S1 Statistical Analysis). Details on variable selection can be found in S1 Statistical Analysis. Individuals nested with homes with homes nested within each of eight communities was set as a random effect. Restricted Maximum likelihood (REML) was used due to the unbalanced nature of our dataset [62]. Abundance of Ae. aegypti females was averaged over 5 weeks as IgG response to bites has been shown to appear between 1 to 6 weeks post exposure [30], and to specifically last for a least 4 weeks in Aedes [64], and to be more persistent than Anopheles IgG responses [65].
We first created a global model (mgloabl1B) that had a Gaussian distribution with an identity link as our outcome variable is continuous. To check model assumptions, we plotted the distribution of residuals and assessed QQ plots. Results showed deviation from the expected distribution, so we next ran another global model (mglobal2B) using a log link for a Gaussian distribution. Plotted residuals from mglobal2 showed a normal distribution. Backward elimination was used to simplify mglobal2B such that simpler models with lower AIC values were kept [57,66].
In addition, we were interested in analyzing the risk factors that modulated Ae. aegypti female abundance, to compare results to out prior study in the same region with independent data [16]. We used a similar approach to our analysis of the Bitemark Assay, using a GLMM to assess different factors that were indicated previously important by a review of the literature. Again, we constructed a global model (mgloabl1A) that included 10 explanatory fixed effects: presence of water storage devices in the yard, total number of containers, host index 1, host index 2, income, vegetation level (i.e., percentage of yard coverage by vegetation) area in m2, orderliness of yard, precipitation, and maximum temperature. Random effects were community and intervention arm. First, we modeled a Poisson error distribution in which variance = mean and compared this to negative binomial distributions (type 1and 2) in which variance > mean [67,68]. Once the correct distribution was chosen based on evaluation of QQ plots and AIC values backward elimination was again used to simplify the models and evaluated based on AIC values. All models were generated, and figures were created using R Version 4.3.2 (September 1, 2023), except for Fig 1 which was created using QGIS (version 3.16.6-Hannover). R code can be found in S1 Code and a more detailed statistical description can be found in S1 Statistical Analysis.
Results
In total, 64 adult humans from different households were interviewed using our KAP survey. The human knowledge level of adult mosquitoes was high with 100% of interviewees recognizing an adult mosquito specimen. Fewer individuals were able to identify larval or pupal mosquitoes (35.9%). Most interviewees (90.6%) believed mosquitoes affect their families either as a nuisance, health risk, or cause of allergies, though the level of the problem they believed mosquitoes caused varied (Table 1). Knowledge of mosquito-borne diseases was also high, 81.3% of respondents had heard of them before the interview. Most participants (82.8%) considered mosquito-borne disease a risk to their community, though few knew someone personally who had been infected with a mosquito-borne pathogen (28.1%). Lots surveyed had an average size of 682 m2 (sd = 236) with a variety of vegetation cover and vegetation height. Few lots (14.3%) actively stored water on their property for later use. However, most lots (76.6%) had other containers that could serve as larval habitat if filled with water via a rain event. Most houses had some form of air conditioning with window units being the most common followed by central systems (Table 2).
Over the course of the five weeks before the blood samples were collected, a total of 1,379 female Ae. aegypti were collected, with an average of 2.9 ± 0.005 caught per trap day. The mean of ΔOD values from the Bitemark Assay for Ae. aegypti was 0.12, with a range of 0.05–0.43. A total of 187 female Ae. albopictus were found during the five weeks prior to blood samples, with an average of 0.26 caught per trap day. No Anopheles were caught during the five weeks prior to human blood sampling. The range of ΔOD for Anopheles was 0.08–0.35 with a mean of 0.18 ± 0.004.
Given the large number of variables we collected, three indices were generated by grouping variables of similar nature together using PCA: door, host, and window. Plots of the PCAs and their interpretation can be found in S1 Statistical Analysis. Descriptive statistics of the explanatory variables can be found in Table A in S1 Statistical Analysis. Minimization of AIC via backward elimination identified the best fit model which considered the following covariates: annual household income, age of participant, lot area, and the female Ae. aegypti abundance per trap night averaged over five weeks prior to human blood sampling. Households with an annual income of >$25,000 (i.e., 35% of households) were 1.21(Exponentiated 95% CI: 1.06–1.39) times more likely to be exposed to Ae. aegypti bites. Age was treated as a continuous variable and was also a significant indicator of bite exposure, with older individuals being 7% less likely to be bitten for each year of age. Larger lots were 1.11(Exponentiated 95% CI: 1.04–1.17) times more likely per m2 to have individuals in the household exposed to bites. For each additional adult female Ae. aegypti in the lot, humans were 1.12 (Exponentiated 95% CI: 1.05–1.18) times more likely to be exposed to bites. Individual effects of each predictor variable can be seen in Fig 2.
aegypti bites. N34kDa indicates antibody levels, income is represented in U.S. dollars, age was measured in years and was standardized, areas measured in m2 and was standardized, and week5Avg is the average number of female Ae. aegypti adults caught per trap night in the 5 weeks prior to sampling.
Additionally, we analyzed factors that were associated with female Ae. aegypti abundance in the yards of sampling households. Evaluation of QQ plots and minimization of AIC values identified the Negative Binomial 2 distribution as the best fit. Backward elimination and assessment of AIC values identified the best fit model (m2A) that included the following variables: water storage in the yard, total number of containers in the yard, host index 2, annual household income, vegetation level in yard, area in m2, and cumulative precipitation in inches. Four of these variables were statistically significant. Unsurprisingly, more water-holding containers in the yard led to more female Ae. aegypti with each additional container leading to 1.16 (Exponentiated 95% CI: 1.06–1.27) increase in mosquitoes. Area had an positive relationship to mosquito abundance, with smaller yards having 18% fewer mosquitoes than larger yard (Exponentiated 95% CI: 0.74–0.90). Higher income (i.e., >$25,000) was associated with lower relative abundance of Ae. aegypti females. And lastly, medium levels of vegetation in the yard had a positive relationship with the number of females in the yard. Individual effects of predictor variables can be seen in Fig 3.
Discussion
The results from our KAP survey indicate that most residents recognized adult mosquitoes, had heard of mosquito-borne disease, and considered it to be a problem in their communities. Even so, less than one third of residents surveyed knew someone personally who had been affected by a mosquito-borne disease. These results further illustrate that the LRGV is an area of low dengue endemicity [10] where human disease outcome variables such as human dengue incidence are not good surveillance or intervention evaluation tools. In this context, measuring human antibody response to species-specific mosquito salivary proteins (i.e., salivary biomarkers) is a good tool to use in areas like the LRGV where infection risk is comparatively low but still present [69].
We evaluated the Bitemark Assay, which measures to IgG response to the Ae. aegypti Nterm-34KDa peptide, to relate female Ae. aegypti abundance to human exposure to their bites with the goal of evaluating a complementary tool in the surveillance of Ae. aegypti, their associated viruses, and interventions aimed at Ae. aegypti population control. Since serological biomarkers have been suggested as a cheaper and quicker option than more traditional entomological measures (e.g., larval or adult abundance indices) [69], it is important to assess the relationship between bite exposure and Ae. aegypti abundance. Our results corroborate a previously established link between Ae. aegypti abundance and bite exposure measured via a serological biomarker, a link that has been shown for both larval [59] and adult [38] abundance measures. Humans in homes with higher abundance of Ae. aegypti in traps had 1.12 times higher exposure for each individual mosquito than humans in homes with fewer Ae. aegypti in traps. We did not sample the indoor vector abundance as we have done in past studies in these communities [46], so we don’t know how this positive but weak association with outdoor abundance would have compared to indoor abundance. In a previous study, a positive relationship has been shown between mosquito density, both indoors and outdoors, and bite exposure [38]. However, several studies on Aedes salivary biomarker’s link to density are either focused on immature mosquito life stages or do not specify the location of trapping, making it difficult to make comparisons to our work [38,44].
Our results indicate that there are other factors–environmental and social in nature—that predict exposure to Ae. aegypti bites in the LRGV. Socially, age and income were significant predictors of exposure. A higher income (i.e., >$25,000), led to more exposure to Ae. aegypti bites, a finding that is an expected extension of Juarez et al. [16] work in the area who found that medium income, i.e., $25,000–50,000, was an indicator of higher outdoor Ae. aegypti relative abundance than in low- or high-income areas. While Martin and colleagues [46] found low-income communities to have a higher relative abundance of Ae. aegypti mosquitoes than mid- or high-income communities, their study design drew income data from the U.S. Census at the block level where our study and Juarez et al. [16] used income data at the household level. These differences in scale could explain the difference in the results, a phenomenon widely described in ecology as the paradox of how resource availability is described depending on a density measurement [70]. Another possible explanation for these differences is that all our communities are classified as low-income and it is possible that higher income households, within this lower income context, have more exposure to Ae. aegypti bites. This could be due to these households have more resources for outdoor spaces such as more plant pots-saucers or water features that could serve as habitat for larva. Moreover, the relationship between socio-economic factors, such as income, and Aedes mosquitoes is variable and often determined by geographical context, with a slight majority (50–60%) of studies showing greater mosquito abundance in areas with lower socio-economic status (SES) [71]. Lastly, it is important to point out that exposure to bites is not a proxy for adult female abundance. Though we observed a positive relationship in between exposure and adult female abundance, our results indicate that they are modulated by different factors and should not be used interchangeably.
To the best of our knowledge, our study is the first to find a relationship between SES, as measured by income, and exposure to Ae. aegypti bites. Other studies have showed mixed results in terms of SES variables and bite exposure. For example, one SES variable, occupation type, has been associated with exposure previously [38] while education level has been shown to have no effect on exposure [60]. Lastly, it is worth noting that the effect of income on exposure to Aedes bites was small but significant, which likely means it explains a small proportion of the variation in the data. Confounding variables, such as mobility, where individuals spend most of the day (e.g., indoors, outdoors, at school), and indoor abundance of Ae. aegypti, that were not directly measured in this study or were not included in the models may be playing a key role. This caveat should also be considered for the effect of age on exposure to Aedes bites.
In our study, age was negatively associated with exposure to Ae. aegypti bites indicating that younger people were more likely to be bitten than older people. In previous studies, age has been a contributor to bite exposure levels, though the directionality of the relationship varied among studies [38,59,60]. Doucoure and colleagues [44] propose three hypotheses to understand the difference in exposure between adults and children: 1) antibody response is directly correlated to the bites received, 2) children have stronger reactions to bites than adults, or 3) adults experience desensitization to bites. Although this study was not designed to interrogate these three hypotheses, we observed that children had different mobility and behaviors than adults, such that they spent most of their time away from home (i.e., at school) or inside (i.e., sleeping). These differences could translate to different exposure levels to Ae. aegypti bites (first hypothesis). It is worth noting that human attractiveness to mosquitoes is highly variable, and can be influenced by factors including skin microbiota, diet, pathogen infection status, and genetics [72]. Specifically, smaller individuals, such as children, may produce less carbon dioxide or volatile chemicals to attract mosquitoes. Our results showed consistency with previous work in regard to lot size [16,73], i.e., larger lots were associated with higher bite exposure. We did not find any association between bite exposure and human mobility, even though previous modelling [74–76] and empirical studies [77] have shown a relationship between mobility and dengue virus transmission. While our results were unexpected, it is important to note that we asked participants about their movement patterns during the interview portion, which could have led to inaccuracies due to memory lapse. Moreover, mobility may affect dengue transmission dynamics as mentioned above, but not bite exposure as Ae. aegypti travels only within a 200m area [47] compared to the kilometers that humans move [72].
Serological biomarkers are valuable tools in a variety of contexts related to arthropod vectors, the pathogens they transmit, and their control. Here we presented a study showing the relationship between the serological biomarker, Aedes Nterm-34kDa peptide, and adult female Ae. aegypti abundance or density as well associated risk factors for bite exposure. An additional application of this tool is the evaluation of vector control interventions. Serological biomarkers, such as gSg6-PI and cE5, have been used to effectively evaluate Anopheles and malaria interventions [31, 32, 78] and the Nterm-34kDa peptide has also been used to effectively evaluate Ae. albopictus [64]. Though, to our knowledge, there has not been use of serological biomarkers to evaluate interventions of Ae. aegypti, this study and another similar study in Thailand [38] were conducted as part of intervention studies that may have yet to be published.
In addition to assessing risk factors related to bite exposure, we investigated how environmental and social risk factors affected adult female Ae. aegypti abundance. The total number of containers in the yard was positively associated with adult female Ae. aegypti abundance. These results seem intuitive, it is important to remember that the type of container can affect its production level [79] as well as the other environmental variables such as water volume, presence of nearby trees, and water temperature [80]. Previously in the LRGV, more tires in a yard led to fewer female mosquitoes in the yard, [16] which may indicate the type of container is an important factor determining productivity in the LRGV. Larger yards and lower annual household income were also associated with more mosquitoes. Lastly, medium levels of vegetation in the yard were associated with more adult mosquitoes.
Our results show that different factors explain the variation in adult mosquito abundance and exposure to their bites. Our best fit model for describing adult abundance showed that the number of containers in the yard, the size of the yard, the vegetation level in the yard, and the household income modulated abundance, but this model only explained 54% of the variation in the data when the random effects were included (conditional R2 = 0.54). So, it is unsurprising that the models do not match well as they explain a small amount of the variation in the data. Moreover, the best fit models for bite exposure and adult abundance were based on different explanatory models which could also explain the differences in the models. These results suggest that serological biomarkers are not a direct proxy for arthropod vector abundance. The two measure represent different things and are modulated by different factors. Given momentum of the field utilizing serological evidence of human and animal exposure to vector bites, more research should investigate how vector abundance relates to vector bites.
Our study has several limitations. First, our group has been working in the LRGV in the same communities for at least 5 years. Because of this, the people we surveyed may have more knowledge about mosquitoes and their biology compared to the wider community since they have had years of exposure to our past community engagement and outreach [81]. Moreover, most of our surveys were conducted during the weekday, meaning that generally retired or individuals without normal business hour work schedules were surveyed. This could have biased the age structure of our study. Additionally, Ae. albopictus are present in the sampling communities and some were caught in our BG traps during the study period. While the Nterm-34kDa serological biomarker is considered specific to the species level [44], cross-reactivity between Ae. aegypti and Ae. albopictus has been observed [45, 64]. Though this could confound the outcomes of our study, Ae. aegypti are more abundant in the area and since Ae. albopictus still pose a risk for Aedes-borne arboviruses, this information is still relevant to the control of their spread.
Our study supports the use of the Nterm-34kDa serological biomarker as a proxy for adult Ae. aegypti entomological surveillance as we found a significant, though weak, relationship between exposure to Ae. aegypti bites and adult female abundance. This relationship is consistent with other studies and in our context, demonstrates serological biomarkers as a valuable tool to use as a proxy for adult female abundance or in areas where transmission of Aedes-borne pathogens is low, such as the LRGV. Additionally, we recommend the use of this tool in risk assessments which can complement more traditional entomological measures as it is a more direct measure of exposure to bites than female abundance alone. In similar studies in the future, we recommend better resolution of human mobility such as the use of GPS units [82] or smart phone apps [83] to improve the relationships of this important variable to vector exposure. And finally, we suggest further studies on the use of the Bitemark Assay to provide insights into the strength of this tool to measure risk factors associated with human exposure to vector bites as well as an outcome variable for vector control trials.
Supporting information
S1 Statistical Analysis. Detailed statistical analysis with R code excerpts included.
https://doi.org/10.1371/journal.pntd.0011953.s001
(DOCX)
Acknowledgments
We appreciate the support of the residents and city and country public health agencies in the Lower Rio Grande Valley, TX who collaborated with us to conduct this study. We thank Danya Garza, Salvador Solis, Odaliz Sauceda, Javier Elizondo, Chris Roundy, and Charlotte Rhodes for their assistance in the field.
References
- 1. Bhatt S, Gething PW, Brady OJ, Messina JP, Farlow AW, Moyes CL, et al. The global distribution and burden of dengue. Nature. 2013 Apr 25;496(7446):504–7. pmid:23563266
- 2.
World Health Organization. Zika Epidemiology Update. 2019.
- 3. Puntasecca CJ, King CH, Labeaud AD. Measuring the global burden of Chikungunya and Zika viruses: A systematic review. Vol. 15, PLoS Neglected Tropical Diseases. Public Library of Science; 2021. pmid:33661908
- 4. Bowman LR, Donegan S, McCall PJ. Is dengue vector control deficient in effectiveness or evidence?: Systematic review and meta-analysis. PLoS Negl Trop Dis. 2016 Mar 17;10(3). pmid:26986468
- 5. Juarez JG, Chaves LF, Garcia-Luna SM, Martin E, Badillo-Vargas I, Medeiros MCI, et al. Variable coverage in an Autocidal Gravid Ovitrap intervention impacts efficacy of Aedes aegypti control. Journal of Applied Ecology. 2021 Jul 13;
- 6. Sharp TM, Lorenzi O, Torres-Velásquez B, Acevedo V, Pérez-Padilla J, Rivera A, et al. Autocidal gravid ovitraps protect humans from chikungunya virus infection by reducing Aedes aegypti mosquito populations. PLoS Negl Trop Dis. 2019 Jul 1;13(7). pmid:31344040
- 7. Utarini A, Indriani C, Ahmad RA, Tantowijoyo W, Arguni E, Ansari MR, et al. Efficacy of Wolbachia-infected mosquito deployments for the control of dengue. New England Journal of Medicine. 2021 Jun 10;384(23):2177–86. pmid:34107180
- 8. Tschampl CA, Undurraga EA, Ledogar RJ, Coloma J, Legorreta-Soberanis J, Paredes-Solís S, et al. Cost-effectiveness of community mobilization (Camino Verde) for dengue prevention in Nicaragua and Mexico: A cluster randomized controlled trial. International Journal of Infectious Diseases. 2020;94:59–67. pmid:32179138
- 9. Kraemer MUG, Reiner RC, Brady OJ, Messina JP, Gilbert M, Pigott DM, et al. Past and future spread of the arbovirus vectors Aedes aegypti and Aedes albopictus. Nat Microbiol. 2019 May 1;4(5):854–63. pmid:30833735
- 10. Olson MF, Juarez JG, Kraemer MUG, Messina JP, Hamer GL. Global patterns of aegyptism without arbovirus. PLoS Negl Trop Dis. 2021 May 1;15(5). pmid:33951038
- 11.
Center for Disease Control and Prevention. Potential Range of Aedes aegypti and Aedes albopictus in the United States, 2017 [Internet]. 2023 [cited 2024 Apr 12]. Available from: https://www.cdc.gov/mosquitoes/mosquito-control/professionals/range.html
- 12.
Texas Department of State Health Services. Health Advisory: Locally Acquired Malaria Case. 2023.
- 13. Eder M, Cortes F, Teixeira de Siqueira Filha N, Araújo de França GV, Degroote S, Braga C, et al. Scoping review on vector-borne diseases in urban areas: Transmission dynamics, vectorial capacity and co-infection. Vol. 7, Infectious Diseases of Poverty. BioMed Central Ltd.; 2018. pmid:30173661
- 14. Hargrove WL, Juárez-Carillo PM, Korc M. Healthy vinton: A health impact assessment focused on water and sanitation in a small rural town on the U.S.-Mexico border. Int J Environ Res Public Health. 2015 Apr 7;12(4):3864–88.
- 15. Braier M. The Right to Light: Visibility and Government in the Rio Grande Valley Colonias. Ann Am Assoc Geogr. 2020 Jul 3;110(4):1208–23.
- 16. Juarez JG, Garcia-Luna S, Medeiros MCI, Dickinson KL, Borucki MK, Frank M, et al. The eco-bio-social factors that modulate Aedes aegypti abundance in south Texas border communities. Insects. 2021;12(2):1–16.
- 17. Reiter P, Lathrop S, Bunning M, Biggerstaff B, Singer D, Tiwari T, et al. Texas lifestyle limits transmission of dengue virus [Internet]. Vol. 9, Emerging Infectious Diseases •. 2003. Available from: http://www.laredochamber.com/contact-
- 18. Vannavong N, Seidu R, Stenström TA, Dada N, Overgaard HJ. Effects of socio-demographic characteristics and household water management on Aedes aegypti production in suburban and rural villages in Laos and Thailand. Parasit Vectors. 2017 Apr 4;10(1). pmid:28376893
- 19. LaDeau SL, Leisnham PT, Biehler D, Bodner D. Higher mosquito production in low-income neighborhoods of baltimore and washington, DC: Understanding ecological drivers and mosquito-borne disease risk in temperate cities. Int J Environ Res Public Health. 2013;10(4):1505–26. pmid:23583963
- 20. Scavo NA, Barrera R, Reyes-Torres LJ, Yee DA. Lower socioeconomic status neighborhoods in Puerto Rico have more diverse mosquito communities and higher Aedes aegypti abundance. Journal of Urban Ecology. 2021;7(1).
- 21. Whiteman A, Gomez C, Rovira J, Chen G, McMillan WO, Loaiza J. Aedes Mosquito Infestation in Socioeconomically Contrasting Neighborhoods of Panama City. Ecohealth. 2019 Jun 15;16(2):210–21. pmid:31114946
- 22. Barrera R, Acevedo V, Amador M. Role of abandoned and vacant houses on Aedes aegypti productivity. American Journal of Tropical Medicine and Hygiene. 2021 Jan 6;104(1):145–50. pmid:33021195
- 23. Paul KK, Dhar-Chowdhury P, Emdad Haque C, Al-Amin HM, Goswami DR, Heel Kafi MA, et al. Risk factors for the presence of dengue vector mosquitoes, and determinants of their prevalence and larval site selection in Dhaka, Bangladesh. PLoS One. 2018 Jun 1;13(6). pmid:29928055
- 24. Nagao Y, Thavara U, Chitnumsup P, Tawatsin A, Chansang C, Campbell-Lendrum D. Climatic and social risk factors for Aedes infestation in rural Thailand. Tropical Medicine and International Health. 2003 Jul 1;8(7):650–9. pmid:12828549
- 25.
World Health Organization. Global Vector Control Response 2017–2030. 2017.
- 26. Mccarthy TA, Hadler JL, Julian K, Walsh SJ, Biggerstaff BJ, Hinten SR, et al. West Nile Virus Serosurvey and Assessment of Personal Prevention Efforts in an Area with Intense Epizootic Activity: Connecticut, 2000.
- 27. McMillan JR, Armstrong PM, Andreadis TG. Patterns of mosquito and arbovirus community composition and ecological indexes of arboviral risk in the Northeast United States. PLoS Negl Trop Dis. 2020 Feb 1;14(2). pmid:32092063
- 28. Rivera A, Adams LE, Sharp TM, Lehman JA, Waterman SH, Paz-Bailey G. Morbidity and Mortality Weekly Report Travel-Associated and Locally Acquired Dengue Cases [Internet]. 2010. Available from: https://www.cdc.gov/dengue/
- 29. Bousema T, Youssef RM, Cook J, Cox J, Alegana VA, Amran J, et al. Serologic markers for detecting malaria in areas of low endemicity, Somalia, 2008. Emerg Infect Dis. 2010 Mar;16(3):392–9. pmid:20202412
- 30. Doucoure S, Drame PM. Salivary biomarkers in the control of mosquito-borne diseases. Vol. 6, Insects. MDPI AG; 2015. p. 961–76. pmid:26593952
- 31. Drame PM, Poinsignon A, Besnard P, Le Mire J, Dos-Santos MA, Sow CS, et al. Human antibody response to Anopheles gambiae saliva: An immuno-epidemiological biomarker to evaluate the efficacy of insecticide-treated nets in malaria vector control. American Journal of Tropical Medicine and Hygiene. 2010 Jul;83(1):115–21. pmid:20595489
- 32. Brosseau L, Drame PM, Besnard P, Toto JC, Foumane V, Le Mire J, et al. Human Antibody Response to Anopheles Saliva for Comparing the Efficacy of Three Malaria Vector Control Methods in Balombo, Angola. PLoS One. 2012 Sep 24;7(9). pmid:23028499
- 33. Londono-Renteria B, Cardenas JC, Cardenas LD, Christofferson RC, Chisenhall DM, Wesson DM, et al. Use of anti-aedes aegypti salivary extract antibody concentration to correlate risk of vector exposure and dengue transmission risk in colombia. PLoS One. 2013 Dec 2;8(12). pmid:24312537
- 34. Montiel J, Carbal LF, Tobón-Castaño A, Vásquez GM, Fisher ML, Londono-Rentería B. IgG antibody response against Anopheles salivary gland proteins in asymptomatic Plasmodium infections in Narino, Colombia. Malar J. 2020 Jan 23;19(1). pmid:31973737
- 35. Londono-Renteria BL, Shakeri H, Rozo-Lopez P, Conway MJ, Duggan N, Jaberi-Douraki M, et al. Serosurvey of Human Antibodies Recognizing Aedes aegypti D7 Salivary Proteins in Colombia. Front Public Health. 2018 May 18;6. pmid:29868532
- 36. Parra MCP, Fávaro EA, Dibo MR, Mondini A, Eiras ÁE, Kroon EG, et al. Using adult Aedes aegypti females to predict areas at risk for dengue transmission: A spatial case-control study. Acta Trop. 2018 Jun 1;182:43–53. pmid:29462598
- 37. Lau SM, Chua TH, Sulaiman WY, Joanne S, Lim YAL, Sekaran SD, et al. A new paradigm for Aedes spp. surveillance using gravid ovipositing sticky trap and NS1 antigen test kit. Parasit Vectors. 2017 Mar 21;10(1). pmid:28327173
- 38. Fustec B, Phanitchat T, Aromseree S, Pientong C, Thaewnongiew K, Ekalaksananan T, et al. Serological biomarker for assessing human exposure to Aedes mosquito bites during a randomized vector control intervention trial in northeastern Thailand. PLoS Negl Trop Dis. 2021;15(5).
- 39. Bazin M, Williams CR. Mosquito traps for urban surveillance: collection efficacy and potential for use by citizen scientists. Journal of Vector Ecology. 2018 Jun 1;43(1):98–103. pmid:29757507
- 40. Londono-Renteria BL, Eisele TP, Keating J, James MA, Wesson DM. Antibody response against Anopheles albimanus (Diptera: Culicidae) salivary protein as a measure of mosquito bite exposure in Haiti. J Med Entomol. 2010 Nov;47(6):1156–63. pmid:21175067
- 41. Londono-Renteria B, Patel JC, Vaughn M, Funkhauser S, Ponnusamy L, Grippin C, et al. Long-lasting permethrin-impregnated clothing protects against mosquito bites in outdoor workers. American Journal of Tropical Medicine and Hygiene. 2015 Oct 1;93(4):869–74. pmid:26195460
- 42. Poinsignon A, Cornelie S, Mestres-Simon M, Lanfrancotti A, Rossignol M, Boulanger D, et al. Novel peptide marker corresponding to salivary protein gSG6 potentially identifies exposure to Anopheles bites. PLoS One. 2008 Jun 25;3(6). pmid:18575604
- 43. Olajiga OM, Maldonado-Ruiz LP, Fatehi S, Cardenas JC, Gonzalez MU, Gutierrez-Silva LY, et al. Association of dengue infection with anti-alpha-gal antibodies, IgM, IgG, IgG1, and IgG2. Front Immunol. 2022 Oct 14;13. pmid:36311743
- 44. Fontaine A, Pascual A, Orlandi-Pradines E, Diouf I, Remoué F, Pagès F, et al. Relationship between exposure to vector bites and antibody responses to mosquito salivary gland extracts. PLoS One. 2011 Dec 14;6(12). pmid:22195000
- 45. Doucoure S, Mouchet F, Cornelie S, DeHecq JS, Rutee AH, Roca Y, et al. Evaluation of the human IgG antibody response to aedes albopictus saliva as a new specific biomarker of exposure to vector bites. PLoS Negl Trop Dis. 2012 Feb;6(2). pmid:22363823
- 46. Martin E, Medeiros MCI, Carbajal E, Valdez E, Juarez JG, Gracia-Luna S, et al. Surveillance of Aedes aegypti indoors and outdoors using Autocidal Gravid Ovitraps in South Texas during local transmission of Zika virus, 2016 to 2018. Acta Trop. 2019 Apr 1;192:129–37. pmid:30763563
- 47. Juarez JG, Garcia-Luna S, Chaves LF, Carbajal E, Valdez E, Avila C, et al. Dispersal of female and male Aedes aegypti from discarded container habitats using a stable isotope mark-capture study design in South Texas. Sci Rep. 2020 Dec 1;10(1). pmid:32321946
- 48.
U.S. Census Bureau. Hidalgo County, Texas [Internet]. 2020 [cited 2019 Mar 14]. Available from: https://www.census.gov/quickfacts/hidalgocountytexas
- 49.
NOAA. Local Climatological Data, McAllen Miller Int Airport, TX. 2017 [cited 2019 Jan 29]. National Weather Service: Climate Prediction Center. Available from: https://w2.weather.gov/climate/xmacis.php?wfo=bro
- 50. Unlu I, Rochlin I, Suman DS, Wang Y, Chandel K, Gaugler R. Large-scale operational pyriproxyfen autodissemination deployment to suppress the immature asian tiger mosquito (Diptera: Culicidae) Populations. J Med Entomol. 2020 Jul 1;57(4):1120–30. pmid:32006427
- 51. Vitek CJ, Gutierrez JA, Dirrigl FJ. Dengue vectors, human activity, and dengue virus transmission potential in the lower rio Grande Valley, Texas, United States. J Med Entomol. 2014;51(5):1019–28. pmid:25276932
- 52. Guerrero CD, Hinojosa S, Vanegas D, Tapangan N, Guajardo M, Alaniz S, et al. Increasing public health mosquito surveillance in hidalgo county, texas to monitor vector and arboviral presence. Pathogens. 2021 Aug 1;10(8). pmid:34451489
- 53.
Darsie R, Ward R. Identification and geographical distribution of the mosquitoes of North America, north of Mexico. University Press of Florida; 2016.
- 54. Manning JE, Oliveira F, Sepulveda N, Londono-Renteria B, Olajiga OM, Marin-Lopez A, et al. Aedes aegypti anti-salivary proteins IgG levels in a cohort of DENV-like symptoms subjects from a dengue-endemic region in Colombia. Frontiers in Epidemiology. 2022;2.
- 55. Chaves LF, Calzada JE, Rigg C, Valderrama A, Gottdenker NL, Saldaña A. Leishmaniasis sand fly vector density reduction is less marked in destitute housing after insecticide thermal fogging. Parasit Vectors. 2013;6(1). pmid:23742709
- 56. Harrison XA, Donaldson L, Correa-Cano ME, Evans J, Fisher DN, Goodwin CED, et al. A brief introduction to mixed effects modelling and multi-model inference in ecology. PeerJ. 2018;2018(5). pmid:29844961
- 57. Burnham KP, Anderson DR. Multimodel inference: Understanding AIC and BIC in model selection. Vol. 33, Sociological Methods and Research. 2004. p. 261–304.
- 58. Ya-Umphan P, Cerqueira D, Parker DM, Cottrell G, Poinsignon A, Remoue F, et al. Use of an anopheles salivary biomarker to assess malaria transmission risk along the Thailand-Myanmar border. Journal of Infectious Diseases. 2017;215(3):396–404. pmid:27932615
- 59. Doucoure S, Mouchet F, Cournil A, Le Goff G, Cornelie S, Roca Y, et al. Human antibody response to Aedes aegypti saliva in an urban population in Bolivia: A new biomarker of exposure to dengue vector bites. American Journal of Tropical Medicine and Hygiene. 2012 Sep;87(3):504–10. pmid:22848099
- 60. Kassam NA, Laswai D, Kulaya N, Kaaya RD, Kajeguka DC, Schmiegelow C, et al. Human IgG responses to Aedes mosquito salivary peptide Nterm-34kDa and its comparison to Anopheles salivary antigen (gSG6-P1) IgG responses measured among individuals living in Lower Moshi, Tanzania. PLoS One. 2022 Oct 1;17(10 October). pmid:36301860
- 61. Calzada JE, Saldaña A, González K, Rigg C, Pineda V, Santamaría AM, et al. Cutaneous Leishmaniasis in dogs: Is high seroprevalence indicative of a reservoir role? Parasitology. 2015 Aug 14;142(9):1202–14. pmid:25990429
- 62. Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR, Stevens MHH, et al. Generalized linear mixed models: a practical guide for ecology and evolution. Vol. 24, Trends in Ecology and Evolution. 2009. p. 127–35. pmid:19185386
- 63. Chaves LF. An entomologist guide to demystify pseudoreplication: Data analysis of field studies with design constraints. J Med Entomol. 2010 May;47(3):291–8. pmid:20496574
- 64. Elanga Ndille E, Doucoure S, Poinsignon A, Mouchet F, Cornelie S, D’Ortenzio E, et al. Human IgG Antibody Response to Aedes Nterm-34kDa Salivary Peptide, an Epidemiological Tool to Assess Vector Control in Chikungunya and Dengue Transmission Area. PLoS Negl Trop Dis. 2016 Dec 1;10(12). pmid:27906987
- 65. Orlandi-Pradines E, Almeras L, Denis de Senneville L, Barbe S, Remoué F, Villard C, et al. Antibody response against saliva antigens of Anopheles gambiae and Aedes aegypti in travellers in tropical Africa. Microbes Infect. 2007 Oct;9(12–13):1454–62. pmid:17913537
- 66.
Venables WN, Ripley BD. Modern Applied Statistics with S. New York City: Springer; 2002.
- 67. Sileshi G. Selecting the right statistical model for analysis of insect count data by using information theoretic measures. Bull Entomol Res. 2006;96(5):479–88. pmid:17092359
- 68. White GC, Bennetts RE. Analysis of frequency count data using the negative binomial distribution. Ecology. 1996;77(8):2549–57.
- 69. Sagna AB, Yobo MC, Ndille EE, Remoue F. New immuno-epidemiological biomarker of human exposure to Aedes vector bites: From concept to applications †. Vol. 3, Tropical Medicine and Infectious Disease. MDPI AG; 2018.
- 70. Lewontin R.C., Levins R. On the characterization of density and resource availability. The American Naturalist. 1989 Oct;134(4).
- 71. Whiteman A, Loaiza JR, Yee DA, Poh KC, Watkins AS, Lucas KJ, et al. Do socioeconomic factors drive Aedes mosquito vectors and their arboviral diseases? A systematic review of dengue, chikungunya, yellow fever, and Zika Virus. One Health. 2020;
- 72. Donnelly MAP, Kluh S, Snyder RE, Barker CM. Quantifying sociodemographic heterogeneities in the distribution of aedes aegypti among California households. PLoS Negl Trop Dis. 2020 Jul 1;14(7):1–21. pmid:32692760
- 73. Ospina-Aguirreid C, Soriano-Paños D, Olivar-Tost G, Galindo-González CC, Gómez-Gardeñes J, Osorio G. Effects of human mobility on the spread of Dengue in the region of Caldas, Colombia. PLoS Negl Trop Dis. 2023 Nov 1;17(11 November).
- 74. Barrios E, Lee S, Vasilieva O. Assessing the effects of daily commuting in two-patch dengue dynamics: A case study of Cali, Colombia. J Theor Biol. 2018 Sep 14;453:14–39. pmid:29775680
- 75. Kiang M V., Santillana M, Chen JT, Onnela JP, Krieger N, Engø-Monsen K, et al. Incorporating human mobility data improves forecasts of Dengue fever in Thailand. Sci Rep. 2021 Dec 1;11(1). pmid:33441598
- 76. Stoddard ST, Morrison AC, Vazquez-Prokopec GM, Soldan VP, Kochel TJ, Kitron U, et al. The role of human movement in the transmission of vector-borne pathogens. PLoS Negl Trop Dis. 2009 Jul;3(7). pmid:19621090
- 77. Ellwanger JH, Cardoso J da C, Chies JAB. Variability in human attractiveness to mosquitoes. Vol. 1, Current Research in Parasitology and Vector-Borne Diseases. Elsevier B.V.; 2021. pmid:35284885
- 78. Marie A, Ronca R, Poinsignon A, Lombardo F, Drame PM, Cornelie S, et al. The Anopheles gambiae cE5 salivary protein: A sensitive biomarker to evaluate the efficacy of insecticide-treated nets in malaria vector control. Microbes Infect. 2015;17(6):409–16. pmid:25637950
- 79. Chadee DD, Huntley S, Focks DA, Chen AA. Aedes aegypti in Jamaica, West Indies: Container productivity profiles to inform control strategies. Tropical Medicine and International Health. 2009 Feb;14(2):220–7. pmid:19236668
- 80. Barrera R, Amador M, Clark GG. Ecological Factors Influencing Aedes aegypti (Diptera: Culicidae) Productivity in Artificial Containers in Salinas, Puerto Rico [Internet]. Vol. 43, POPULATION AND COMMUNITY ECOLOGY J. Med. Entomol. 2006. Available from: https://academic.oup.com/jme/article/43/3/484/879416
- 81. Juarez JG, Carbajal E, Dickinson KL, Garcia-Luna S, Vuong N, Mutebi JP, et al. The unreachable doorbells of South Texas: community engagement in colonias on the US-Mexico border for mosquito control. BMC Public Health. 2022 Dec 1;22(1). pmid:35698216
- 82. Schaber KL, Paz-Soldan VA, Morrison AC, Elson WHD, Rothman AL, Mores CN, et al. Dengue illness impacts daily human mobility patterns in Iquitos, Peru. PLoS Negl Trop Dis. 2019;13(9). pmid:31545804
- 83. Kache PA, Bron GM, Zapata-Ramirez S, Tsao JI, Bartholomay LC, Paskewitz SM, et al. Evaluating spatial and temporal patterns of tick exposure in the United States using community science data submitted through a smartphone application. Ticks Tick Borne Dis. 2023 Jul 1;14(4). pmid:37001417