Perspectives and Challenges in Entomological Risk Assessment and Vector Control of Chikungunya

Abstract

Originally identified in Tanzania in 1953 [1], chikungunya virus (CHIKV), a mosquito-borne alphavirus, caused minor outbreaks in Africa, India, and Southeast Asia for several decades [2]. In 2004, a CHIKV strain of East, Central, and Southern African lineage led to a large outbreak in Kenya that quickly spread to several Indian Ocean islands and India, causing explosive epidemics with eventual dissemination to other parts of Asia and Europe [2–4]. In 2013, an Asian lineage CHIKV strain was introduced in St Martin and quickly spread throughout the Americas [5]. The epidemic proportions of the spread of CHIKV are due to several factors: increased urbanization; previously naive human populations; the invasiveness and widespread geographic distribution of a major mosquito vector, Aedes albopictus; and adaptive mutations in the Indian Ocean lineage CHIKV, which resulted in enhanced transmission and vector competence in A. albopictus [6–8]. Because there is no commercial vaccine or effective treatments for CHIKV [4], vector control and prevention remain the best strategies to minimize the brunt of CHIKV transmission. We will provide an overview of entomological risk assessment and vector control methods for CHIKV.

CHIKV MOSQUITO VECTORS

A. aegypti and A. albopictus are the 2 main CHIKV mosquito vectors; however, numerous mosquito species have been observed to transmit the virus (Table 1). Both A. aegypti and A. albopictus are forest species that exploit small, natural containers of water (eg, holes in trees, fruit husks, and plant axils) for their aquatic, developmental stages. They have also adapted to the urban environment, where they colonize a variety of artificial water containers (eg, buckets, broken crockery, flowerpot saucers, flower vases, derelict cars, water storage drums, and tires). A. aegypti is highly anthropophilic and is a primarily endophilic (indoor) and a daytime and crepuscular biter [9, 10]. A. albopictus generally lives in rural and peridomestic areas and is implicated in the maintenance of the CHIKV sylvatic cycle. It feeds on many animals and it is mainly a diurnal, outdoor biter with partial endophilic and endophagous behaviors. A. albopictus mainly breeds in outdoor storage and discarded containers. A highly invasive mosquito species, it has greatly expanded into tropical and temperate areas in recent years [11]. Studies have also shown A. albopictus to be highly urbanized, anthropophilic, and implicated in CHIKV outbreaks [12–14].

ENTOMOLOGICAL INVESTIGATIONS TO ASSESS THE RISK OF CHIKV AMPLIFICATION

Vector surveillance and control are used to monitor, reduce, and maintain vector density at a level that does not allow epidemic transmission. However, this objective and charge is difficult to implement, partly because of the lack of a correlation between epidemic transmission and vector density. For example, an area can have high vector density without viral presence, and an epidemic can be supported by low vector density. This complicates the entomological risk assessment for amplification of arthropod viruses (arboviruses). Furthermore, sylvatic CHIKV, maintained predominantly in nonhuman primate reservoirs, has the potential to cause epidemics, and very little is known about the forces that govern the sylvatic cycle. Recent work has highlighted the fact that Aedes abundance and climate affect arboviral circulation [15]. In addition, a study on the temporal and spatial dynamics of sylvatic CHIKV across varying landscapes in Senegal implicated Aedes furcifer as the vector most responsible for CHIKV spillover in humans owing to its broad land coverage [16] (Figure 1). Such studies begin to move the field forward, clarifying the intricate interactions among the environment, animal reservoirs, the virus, and the mosquito, which all collectively determine the risk of CHIKV outbreaks.

Figure 1.

Chikungunya virus (CHIKV)–infected Aedes species found in varying landscapes during the 2009 amplification event in Kédougou, Senegal. Colors indicate landscape class in which Aedes mosquitoes were collected. Values in parentheses represent number of CHIKV-positive landscapes; the numbers investigated were equal for all landscapes. Abbreviation: UTM, Universal Transverse Mercator coordinate system.

Open in new tabDownload slide

Although CHIKV and other Aedes-transmitted viruses, such as dengue virus (DENV), yellow fever virus (YFV), and Zika virus differ from each other, there are no standard vector control strategies for them individually and no single entomological risk assessment tool employed to discern viral amplification. These diseases share the same epidemic vectors; A. aegypti transmits all 4 diseases, and A. albopictus transmits DENV and CHIKV.

Assessing the risk of viral amplification in A. aegypti and A. albopictus will provide insight into the risk of disease outbreaks, and such knowledge can then be used to inform decisions on vector control and surveillance programs. The overall objective of entomological risk assessment is to collect enough data on vector bionomics to determine the spatiotemporal risks of viral amplification. Specifically, the objectives of an entomological risk assessment are to identify the vectors, determine their distribution and abundance (by collecting adults or indirectly by using immature stages as proxies of adult density), and, finally, estimating their entomological infection and inoculation rates. The following sections detail the list of steps needed to achieve such goals.

Selection of Sampling Sites

The selection of sampling sites is crucial and depends on whether the context is epidemic or not. Priority is given to areas where cases have been reported or to areas visited to encompass the source of infection or zone of extension. Areas where viral antigens are detected are also investigated to facilitate virus detection and vector identification. However, the tendency to assess the risk of infection in known affected areas is questionable because local communities may develop protective immunity, thereby significantly reducing the probability of viral emergence. To overcome this problem, an approach that uses a geographic information system and remote sensing may identify ecosystems favorable to mosquito vector development, and a profile of disease transmission in different climatic and geographic zones may thereby be developed to gain a more holistic view of outbreak risk. Distinct ecological zones are identified based on environmental and climatic parameters (rainfall, vegetation, temperature, and humidity).The random selection of uniform ecological study sites makes it possible to extrapolate data. The process of sampling site selection follows 3 steps: (1) the mapping of ecological zones using field data or satellite images, (2) the random selection of sites within each zone, and (3) the choice of urban, rural, and forest areas for each site in order to investigate both domestic and sylvatic vectors [17] (Figure 2).

Figure 2.

Sampling site selection process to assess the chikungunya virus entomological risk in sampling zones 1, 2, and 3.

Open in new tabDownload slide

Sampling Mosquito Populations

Currently used approaches explore both the immature developmental stages (larvae or pupae) and the adults. Adult mosquitoes are known to be useful to census the vectors and estimate the population directly involved in transmission and detect viral activity. The immature stages can help to estimate risks according to available indices and also guide decision making for vector control strategies. In certain circumstances, the collection of eggs may be necessary. Sampling must be performed at randomly selected sites, and surveys should be conducted at each site, both indoors and outdoors. In addition, periodic sampling during the rainy season is crucial to identifying all vectors present in the area and accurately estimating the risk of CHIKV amplification.

Larval and Pupae Collections

The immature stages are used as a proxy to estimate the density of the adult populations. This is done by door-to-door inspection of all water containers in and around selected premises, to collect and estimate the numbers of associated larvae or pupae, along with species identification. The main difficulty here is the identification of all positive breeding sites. This task is laborious and not very sensitive, and the results are strongly dependent on the investigators' efforts as well as the consent of the homeowners.

Once obtained, data can be used to calculate epidemic risk indices and to draw up a vector habitat profile, thereby providing a practical way to guide vector control actions. Several indices have been developed to estimate the risk of disease outbreaks [18], which are used more frequently for DENV and YFV. The Breteau index (number of containers found positive for larvae and pupae per 100 houses surveyed), container index (percentage of containers found positive), pupa index (number of pupae per 100 houses inspected), pupae per person index (number of pupae per person), and house index (percentage of houses found positive for larvae or pupae) are commonly used. In addition, various other indices are available (eg, the larval index [number of larvae per 100 houses], larval-pupae index [total number of immature stages per 100 houses], and the productivity of a container [number of the third and fourth mosquito lavae stages {L3–4} or pupae in each container type divided by the total number of L3–4 or pupae in all container types]), and these are based on the presence or absence of larvae or pupae as well as their specific numbers [19]. Although these current tools are useful for estimating epidemic risk, they have serious shortcomings that need to be addressed.

The Breteau, house, and container indices do not take into consideration breeding site productivity (number of immature individuals) but rather consider only the presence or absence of larvae or pupae. The pupae indices ignore larvae, which have generally higher density than pupae. The ultimate objective in the aquatic stage assessment is to determine the most effective control methods. Thus, it is risky to count all containers or houses positive for larvae as negative just because no pupae are present. Furthermore, although larval surveys record containers only as positive or negative, the rate of emergence of adult mosquitoes may vary from one container type to another (eg, bottles vs barrels). Thus, adult abundance, and therefore potential transmission, could be very different in areas with similar larval indices but different container profiles. Moreover, larval indices are poor proxies of aggressive female densities. Not all immature stages will emerge into adults, and only about 50% of these adults will be females and hence involved in transmission. Finally, these indices have not yet been established for assessing the risk of CHIKV amplification. The fact that there is currently no CHIKV amplification threshold for these indices is a huge gap in need of active scientific investigation.

Egg Collections

Eggs are collected using ovitraps (for an example see [20, 21]), typically made in black pots containing water-soluble substrates and attractants for gravid females. These traps provide the advantage of detecting adults during periods of low vector activity. However, their limitations are the lack of overall attraction compared with natural breeding sites, their interference with water containers present in the areas of interest, and an aspect of vector behavior known as oviposition skipping, which may cause a bias in vector density [22].

Adult Mosquito Collections

Adult mosquito sampling looks at 3 types of female mosquitoes: those that are seeking a host, those that are resting, and those that are gravid. For each sampling strategy, different traps have been proposed, with the main objective to census the vector and estimate its density, infection, and entomological inoculation rate (EIR). The EIR measures the intensity of viral transmission (infectious bites per unit of time) and is a direct reflection of vector control efficiency. Human landing collections (HLCs) are performed to collect host-seeking mosquitoes and estimate the EIR [23, 24]. This is the only effective method for estimating host-vector interactions and for collecting both domestic and sylvatic host-seeking CHIKV vectors. HLCs are not usually recommended, especially during epidemics, because of ethical considerations. Biases due to difference in skills and mosquito attraction to collectors and the selective sampling of only unfed females seeking hosts for a blood meal are among other key disadvantages of HLCs. Several alternative methods have been used to collect CHIKV vectors. The BG-Sentinel trap, which is baited with lures (eg, carbon dioxide, octenol, and BG-Lure dispenser), is effective in capturing all population types (unfed, engorged, and gravid) of A. aegypti adult females [25, 26] and A. albopictus [27]. The BG-Sentinel trap uses a combination of attractive visual cues and convection currents that mimic those created by the human body, thus closely resembling HLCs. It also has the advantage of being collapsible, light, and operable all day long [28]. However, data are lacking about the effectiveness of BG-Sentinel traps in collecting sylvatic vector species.

For resting mosquitoes, several kinds of mechanical aspirators used for both indoor and outdoor collections have been proposed. The most commonly used trap is the Centers for Disease Control and Prevention's backpack aspirator [29], which collects all females (unfed, engorged, and gravid) and males [30]. However, sampling resting mosquito populations is intrusive, difficult to standardize (especially for outdoor resting sites), labor intensive, and time consuming. In addition, it requires trained personnel and sufficient sample sizes. Moreover, the resting sites of some species are unknown or not easily accessible [31]. Pyrethrum spray collection is an easy and efficient method to collect indoor resting mosquitoes, does not require specialized equipment, and can be performed by less qualified personnel [24]. However, this method misses all of the outdoor resting mosquitoes. This method is not efficient for virus detection because only dead specimens are collected, and the insecticide used is potentially toxic to the virus.

Oviposition traps (sticky traps and gravid traps) use water and lures as attractants and funnels or sticky boards to capture A. aegypti and A. albopictus gravid females and even males [20, 21, 32]. Gravid traps are highly effective in sampling virus-positive females (as detected by reverse-transcription polymerase chain reaction) and are considerably cheaper and easier to operate. However, sticky traps are not useful for virus isolation because they collect dead specimens. Both techniques miss host-seeking females and sylvatic species breeding in rural habitats. They are also not useful for comparisons across various collection areas and environments because of the differential availability of breeding sites.

CHIKV VECTOR CONTROL

Vector control is critical to curb transmission and prevent outbreaks. The historical ebb and current flow of arboviral vector control (eg, DENV and YFV) can inform us of how best to control CHIKV. In the early 20th century, strictly managed “clean-up campaigns” involving the systematic elimination of breeding sites brought about a remarkable reduction in urban YFV. However, this “source reduction” was highly labor intensive, and in the 1950s, DDT was used to spray infested containers and solid surfaces in a 50-cm radius. This “per-focal treatment” was so effective that the Pan American Health Organization organized a campaign to eradicate A. aegypti from all of the Americas. By the early 1960s, the species was declared absent from 22 countries. Unfortunately, for a variety of reasons, the campaign was abandoned, and A. aegypti reemerged, quickly followed by the reappearance of epidemic DENV [33].

Since this eradication campaign, no country has ever achieved anything close to sustained control of DENV, YFV, or CHIKV. In Singapore, long considered the reference standard in mosquito vector control, DENV epidemics continue despite an annual control budget of $30 million [34]. The invasiveness and enhanced spread of Aedes, due to commercial trade and desiccation-resistant eggs, has led to the persistence of arboviruses [10, 35, 36]. In addition, the widespread use of insecticides has led to increased resistance in vector populations, eliciting the need for better monitoring and the development of alternative strategies [37]. In this section, we discuss the current vector strategies targeting immature and adult mosquito populations.

Control of Immature Populations

During the past 30 years, the World Health Organization and many public health authorities have promoted Communication for Behavioral Impact as a key strategy for DENV control. Its aim is to use education campaigns to encourage the reduction or elimination of infested containers. Unfortunately, there is no convincing evidence of its sustained impact on transmission anywhere in the world, and the high number and diversity of breeding sites makes it logistically difficult to destroy or modify them. Source reduction via the application of toxic, chemical products to aquatic stages or the use of biological agents is considered a more viable and immediate vector control strategy [38]. Among the multitude of insecticide products available, the water-soluble organophosphate larvicide temephos is the most popular, and it can be used alone or in combination with other preventable measures [39]. Despite the popularity of this agent for several decades, many households refuse to treat their water because of assumed risks to humans, and the actual concentration of temephos varies according to the volume of treated water. Another important issue is the depth of water in storage containers, which may result in a concentration gradient of temephos that decreases toward the surface of the water where larvae feed [39].

Aside from toxic chemical agents, plant-based oils [40] and other bioagents function as Aedes larvicides. Examples include the use of copepods and larvivorous fish in communities endemic for arboviruses as well as the use of the larvicidal bacteria Bacillus thuringiensis var. israelensis [41, 42]. It is generally believed that combined source reduction methods work best. However, the implementation of such combination methods poses several challenges, including inconsistent application, diversity and the vast amount of breeding sites, and the need for community engagement and participation; the success of such strategies definitely depends on compliance in the community.

Control of Adult Populations

Historically, the control of adult mosquito populations was based on the use of insecticides alone or insecticide-treated physical barriers (eg, bed nets and curtains). Most communities prefer to invest in the dispersal of aerosolized insecticides indoors and outdoors. Although an evident impact on indoor mosquito populations has been observed [43], the results are less conclusive in peridomestic environments [44]. Despite observed high adult mortality rates, insecticide resistance and adverse impacts on the environment and human health are problematic [45]. To curb these negative effects, new and environmentally friendly strategies are currently being investigated and implemented in the field. For example, Wolbachia, an endosymbiotic bacteria, has been shown to inhibit CHIKV replication in A. aegypti and A. albopictus, and it effectively spreads in the environment according to recent field experiments [46–48]. In another approach, the release of sterilized transgenic male mosquitoes (genetically modified) possessing lethal genes has been shown to reduce Aedes adult populations [49, 50]. Thus, there is a real impact on the vectors in the field, highlighting the feasibility of these technologies on a larger scale.

GAPS AND PERSPECTIVES

To diminish CHIKV as a public health threat, accurate measurements of entomological risk assessment and effective, sustainable vector control strategies are needed. However, to achieve such goals, a sustainable, skilled, diligent workforce is needed, along with a collaborative network of many scientific disciplines (eg, entomologists, virologists, and epidemiologists) to provide high-quality surveillance.

In particular, mosquito collection tools should be standardized to fully capture mosquito-CHIKV interactions. This in turn will allow for more accurate determination of the EIR, which reflects the efficacy of vector control programs. Moreover, the accuracy of the EIR is affected not only by collection methods but also by the skill and diligence of collection teams. To date, no traps are fully efficient at collecting domestic and sylvatic CHIKV vectors, and many mosquito species are associated with CHIKV in nature, especially in enzootic foci [16] (Table 1). In addition, few data are available on the vector competence of these sylvatic vectors. Currently, no mosquito collection tools have been systematically assessed with respect to CHIKV transmission risk. It is also important to highlight the absence of a correlation between vector density, infection, EIRs, and CHIKV prevalence in humans. The diversity of sampling tools makes the comparison between various geographic contexts difficult. In fact, most risk assessment activities are confined to human environments, ignoring the potential existence and importance of the sylvatic cycle. Promoting an approach to identify ecosystems most favorable for vector development and virus transmission is absolutely crucial. Thus, there is an impetus to develop new, more efficient sampling tools that are able to collect all mosquito populations and species involved in CHIKV transmission. Available risk indices (Breteau, house, container, pupae, and pupae per person) should be used to assess CHIKV transmission and surveillance programs in the meantime. However, other indices more specific to CHIKV transmission should be quickly developed, evaluated, and validated.

There is also a need to define epidemic risk thresholds for entomological parameters (eg, vector density, spatiotemporal dynamics, insecticide resistance, and EIR) estimated during entomological risk assessment. Accurate entomological risk assessment should also be done before releasing genetically modified mosquitoes to guide effective vector control programs based on such technologies without wasting resources because of lacking entomological or socioanthropological data. Finally, investing more resources and funds to build sustainable vector management programs is essential to controlling and preventing CHIKV outbreaks. However, the key to the success of vector control programs lies not only in accurate entomological risk assessments and a skilled workforce but also in engaged and educated communities.

Notes

Acknowledgments. We would like to express our deep gratitude to Dr Nancy A. Touchette and Dr Ann M. Powers for their invitation to participate in this chikungunya supplement and for their kind help in editing the manuscript.

Author contributions. M. D., I. D., D. D., S. Y. conceived and designed the manuscript. C. T. D. carried out the map. All authors contributed to the manuscript writing and read, critically revised, and approved the final manuscript.

Financial support. This work was supported by the Institut Pasteur de Dakar.

Potential conflicts of interest. All authors: No potential conflicts of interest. All authors have submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Conflicts that the editors consider relevant to the content of the manuscript have been disclosed.

References

Author notes