Enhanced vector surveillance to control arbovirus epidemics in Colombia

ABSTRACT In the wake of the Zika epidemic, there has been intensified interest in the surveillance and control of the arbovirus vectors Aedes aegypti and Aedes albopictus, yet many existing surveillance systems could benefit from improvements. Vector control programs are often directed by national governments, but are carried out at the local level, resulting in the discounting of spatial heterogeneities in ecology and epidemiology. Furthermore, entomological and epidemiological data are often collected by separate governmental entities, which can slow vector control responses to outbreaks. Colombia has adopted several approaches to address these issues. First, a web-based, georeferenced Aedes surveillance system called SIVIEN AEDES was developed to allow field entomologists to record vector abundance and insecticide resistance data. Second, autocidal gravid oviposition (AGO) traps are deployed as an alternative way to measure vector abundance. Third, data collected by SIVIEN AEDES are used to develop mathematical models predicting Ae. aegypti abundance down to a city block, thus allowing public health authorities to target interventions to specific neighborhoods within cities. Finally, insecticide resistance is monitored through bioassays and molecular testing in 15 high-priority cities, providing a comprehensive basis to inform decisions about insecticide use in different regions. The next step will be to synchronize SIVIEN AEDES data together with epidemiological and climatic data to improve the understanding of the drivers of local variations in arbovirus transmission dynamics. By integrating these surveillance data, health authorities will be better equipped to develop tailored and timely solutions to control and prevent Aedes-borne arbovirus outbreaks.

The urban/peri-urban mosquitoes Aedes aegypti and Aedes albopictus are responsible for transmitting many arboviruses, including Zika, dengue, chikungunya, and yellow fever. Since the 2015 emergence of Zika in Brazil and associated cases of microcephaly (1), there has been intensified interest in mosquito surveillance and control. Growing concerns about potential emergent arboviruses, in combination with the continued geographic expansion of Ae. aegypti and Ae. albopictus, point to the need for increased investment in vector surveillance. Given the lack of effective chemotherapeutics and suitable vaccines for dengue, Zika, and chikungunya, arbovirus control programs rely heavily on the suppression of vector populations to prevent epidemics and control the spread of outbreaks.
Both Ae. aegypti and Ae. albopictus are invasive in the Americas. Ae. aegypti was introduced from Africa via trade ships in the seventeenth through nineteenth centuries, and Ae. albopictus was introduced more recently via the used tire trade in the 1970s-1980s (2,3). In the mid-twentieth century, Ae. aegypti was nearly eradicated as a result of an aggressive Pan American Health Organization yellow fever control program, but the program was abandoned as yellow fever cases waned, resulting in the recovery of mosquito populations. Since that time, Ae. aegypti and Ae. albopictus populations have persisted in the Americas, with the former mosquito generally dominating in highly urbanized areas and the latter in peri-urban and suburban areas.
Today, vector control programs exploit the anthropophillic nature of these mosquitoes by eliminating artificial water-holding containers that may serve as potential larval habitats and by applying larvicides, residual insecticides, and space sprays of adulticides in and around homes. Although there is some evidence that these interventions can be effective, vector control programs can fall short in two important ways. First, vector control programs sometimes assume a top-down structure in which central governments dictate the interventions applied in counties or cities. As a result, important spatial heterogeneities in ecology, entomology, and epidemiology can be overlooked. Second, epidemiological and entomological data are traditionally collected independently by different governmental entities and are only later merged for analyses. This division between vector data and human data makes it difficult to respond quickly to an outbreak and to disentangle the relationships among mosquito population dynamics, vector control interventions, ecological conditions, and human cases of disease. Indeed, these relationships are currently not well characterized or understood (4), and, consequently, interventions are sometimes driven by political pressure rather than empirical evidence.
Here, we describe an initiative led by Colombia's National Institute of Health (INS) that addresses the need to account for spatial heterogeneity in vector control programs, while simultaneously setting the stage for combining entomological and epidemiological data into the same system. We also report on a more comprehensive approach to monitoring insecticide resistance, which is a fundamental component of the architecture of arbovirus surveillance and control programs.

ENHANCING VECTOR SURVEILLANCE IN COLOMBIA: AN INTEGRATED APPROACH
The Colombian National Entomology Network is a unit within the INS that is tasked with developing vector distribution maps, providing reference laboratory services, and compiling data on vector surveillance and control activities as reported by field entomologists from each of the country's 32 departments. As of 2017, Ae. aegypti was found in all 32 departments of Colombia, and Ae. albopictus had been reported in 11 of the 32 (34%), with geographic distribution highly dependent on subdepartmental factors such as access to oviposition sites, human population density, temperature, and altitude below 2 200 m. Data exist at coarse scales, but detailed, fine-resolution information about these covariates, in addition to mosquito presence and abundance data, would allow public health authorities to more effectively tailor control activities to mitigate local risk factors. To address this discrepancy, in 2016 the INS collaborated with the U.S. Centers for Disease Control and Prevention (CDC) to develop plans to: 1) implement an enhanced entomological surveillance system for arbovirus vectors; 2) trial autocidal gravid oviposition traps (AGO traps) to improve measures of vector abundance; 3) model Ae. aegypti abundance using baseline entomological data to determine neighborhoods and city blocks at greatest risk; and 4) comprehensively monitor insecticide resistance trends in selected cities.
A web-based surveillance system for malaria vectors (Sistema de Información para la Vigilancia Entomológica (SIVIEN) MALARIA) is used by INS for compiling data and generating reports related to malaria transmission and control, yet no parallel system existed for Ae. aegypti or Ae. albopictus, despite their public health importance and widespread distribution in Colombia. A similar web-based system for arbovirus vectors, SIVIEN AEDES, has now been developed, and can combine vector data into a nationally accessible geographic information system (GIS) using the QGIS software program (http:// qgis.osgeo.org). Field entomologists throughout Colombia can access SIVIEN AEDES to record information on Aedes insecticide resistance (described below) and abundance (measured by both immature and adult indices). They can also report the discovery of these species in new geographical areas. An associated mobile data collection app called Aedes INSpector allows SIVIEN AEDES to compile data in real time and facilitates the prompt generation of statistical reports. This is useful for rapidly detecting spatiotemporal changes in vector distribution as well as for monitoring and evaluating vector control programs.
Aedes surveillance commonly relies on measurements of immature mosquitoes, obtained either via container surveys or ovitraps. To determine whether improved information can be gained through novel measures of adult mosquito abundance, a pilot study using autocidal gravid oviposition traps (5) (AGO traps) was undertaken in five cities in Colombia. Data obtained from these traps are being compared with the data obtained via traditional immature surveys. Qualitative evaluations are also being carried out to assess the acceptability of AGO traps in households and to evaluate the ease of processing material collected from the traps.
When considered together with epidemiological information, entomological data can be used to guide when and how to mount an elevated vector control response to arbovirus outbreaks. In the same subset of five cities in which AGO traps were installed, mathematical models are being developed in the MATLAB computer program to predict Ae. aegypti abundance at each life stage. The models are adaptive in nature: data on the number of eggs, larvae, pupae, and adult mosquitoes are collected weekly and are continually input into differential equations to evaluate mosquito population dynamics. Results are plotted on heat maps, displaying the approximate number of mosquitoes per city block where the AGO traps were placed. Local authorities can then utilize this information and target additional vector control activities to the specific city blocks with the greatest abundance of mosquitoes. If successful, the INS will expand this approach to monitor mosquito populations in other areas of arbovirus risk.
Insecticide resistance in Colombia is of increasing concern, and data from 2004 through 2015 indicate increasing resistance to pyrethroids among Ae. aegypti populations. The main insecticides used in Colombia for arbovirus control include the organophosphates malathion and pirimiphos-methyl, and the pyrethroids deltamethrin and lambda-cyhalothrin. Although pyrethroid resistance has been linked to the kdr Il,016 substitution on the voltage-gated sodium channel gene on the north coast of Colombia (6), there is a need to characterize other mechanisms of resistance, as well as the intensity of resistance and its impact on the success of vector control interventions throughout the country. Accordingly, the intensity of resistance is being measured through bioassays in 15 cities, and these results are being compared with historical information on vector control interventions. Molecular testing is being conducted to determine the presence of additional kdr mutations (Leu410, Iso1016, and Cys1534), in addition to biochemical tests to determine metabolic mechanisms of resistance. To inform decision-making about which insecticides should be applied in different regions, maps of resistance patterns and their underlying mechanisms are being developed.

CONCLUSION
Zika, dengue, chikungunya, and yellow fever are arboviruses of major concern in Colombia, and robust vector surveillance and control programs are necessary to prevent disease transmission. The enhanced surveillance activities we describe here provide a timely example of how entomological data can be collected through innovative digital platforms, with the aim of improving vector control decision-making at a local scale. Colombia's INS is also planning to synchronize entomological data (collected through SIVIEN AEDES) and epidemiological data (collected through the country's Public Health Surveillance System (SIVIGILA) (https://www.minsalud.gov.co/salud/ Paginas/SIVIGILA.aspx)). Ultimately, combining epidemiological and entomological data will help elucidate patterns that both improve scientific understanding of local transmission dynamics and guide decisions about when and how to scale up interventions in response to outbreaks.
Author contributions. All the authors conceived the original ideas. JMCA and JDRG developed the mathematical models. SJG, AL, and RSL wrote the paper. All the authors reviewed and approved the final version.

Conflicts of interests. None declared.
Funding. This publication was made possible through support provided by the Office of Infectious Disease, Bureau for Global Health, U.S. Agency for International Development, under the terms of an Interagency Agreement with CDC. The opinions expressed herein are those of the authors and do not necessarily reflect the views of the U.S. Agency for International Development.
Disclaimer. The findings and conclusions in this paper are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. In addition, the authors hold sole responsibility for the views expressed in the manuscript, which may not necessarily reflect the opinion or policy of the RPSP/PAJPH or the Pan American Health Organization (PAHO).