Evaluation of an ivermectin-based attractive targeted sugar bait (ATSB) against Aedes aegypti in Tanzania.

Background The control of vector borne arboviral diseases such as Dengue is mainly achieved by reducing human-vector contact and controlling the vectors through source reduction and environmental management. These measures are constrained by labour intensity, insecticide resistance and pro-active community participation. The current study intended to develop and test an ivermectin-based attractive-targeted sugar bait (ATSB) against Aedes aegypti. Methods The 48hour lethal concentration (LC90) of ivermectin against Ae. aegypti was determined through serial dilution experiment where five 30cm x 30cm x 30cm cages were set; into each, a 10% sugar solution treated with ivermectin were introduced. 40 Ae. aegypti were released into each cage and observed for mortality after 4, 8, 24 and 48 hours. The ivermectin-based ATSB was evaluated in a semi field system where ATSB and attractive sugar bait (ASB) were deployed into each compartment of the semi field and 100 female Ae. aegypti were released every day and recaptured the next day through human land catch and Bio-gent sentinel trap. The developed and semi-field tested ATSB was further tested in the field by deploying them in garages. Results The ivermectin 48hr LC90 of male and female Ae. aegypti was found to be 0.03% w/v. In the semi field system, the ATSB significantly reduced a free-flying population of Ae. aegypti within 24 hours (incidence rate ratio (IRR) = 0.62; [95% confidence interval (95%CI); 0.54-0.70] and p-value < 0.001). However, in the field, the ATSBs required the addition of yeast as a carbon dioxide source to efficiently attract Ae. aegypti mosquitoes to feed. Conclusion Ivermectin is an active ingredient that can be used in an ATSB for Ae. aegypti depopulation. However, further research is needed to improve the developed and tested ATSB to compete with natural sources of sugar in a natural environment.


Introduction
Dengue fever is a global public health concern estimated to threaten 100 to 200 million people per year worldwide, with 2.5 billion people worldwide approximated to be at risk (World Health Organization et al., 2009). The disease is endemic in 100 countries, with Asian and South American countries being the most affected (Kyle & Harris, 2008). The disease is reported to have increased by 30-fold in the past half century (World Health Organization, 2012) and is projected to increase further in the coming years (World Health Organization [WHO], 2020a). Reports from Bhatt et al., 2013 have shown that the disease is spreading globally and can no longer be viewed as a regional problem, with cases occurring in countries without past history of disease outbreaks ( . The intervention exploits mosquito sugar feeding behaviour, which is common to both male and female mosquitoes. Usually male mosquitoes exclusively feed on sugar for their whole life while females feed on sugar for survival, flight and fecundity enhancement (Foster, 1995) but require blood for egg-laying. Therefore, an intervention that targets this behaviour will concurrently target both male and female mosquitoes. ATSBs (Foy et al., 2011). Ivermectin also has a good safety record for use in humans and other mammals making it suitable for a bait solution in case of accidental ingestion. For more than three decades, it has been distributed in mass drug administration (MDA) campaigns against filariasis and onchocerciasis with over one billion doses donated by the Mectizan project (Crump & Omura, 2011). The safety of the endectocide is likely related to its target receptors in arthropods where upon being ingested, the ivermectin directly targets glutamate-gated chloride channels (GluCl), a receptor channel which exists in invertebrate nerve and muscle cells leading to paralysis and death of the insect (Fox, 2006;Sheriff et al., 2005) but is non-existent in vertebrates. This study aimed at evaluating the efficacy of an ivermectin-based ATSB as a potential approach for controlling Ae. aegypti mosquitoes in urban Tanzania.

Mosquitoes
All experiments were conducted at Ifakara Health Institute Bagamoyo and Ifakara branches using disease free, insectary-reared Aedes aegypti. The mosquitoes were reared at 27±5 0 C and 40%-99% humidity at Ifakara Health Institute insectary in Bagamoyo and Ifakara, Tanzania. Larvae were fed daily with Tetramin ® fish food while adults were maintained with 10% w/v sugar solution ad libitum with natural light regimen. For egg laying, they were fed cow blood through a membrane.

Laboratory experiment
Determination of ivermectin 48hr LC90 for Ae. aegypti. The dose of ivermectin sufficient to kill 90% of Ae. aegypti was determined in a laboratory using insectary-reared mosquitoes. Serial dilutions of ivermectin in 10% w/v sugar solution were obtained starting with 0.01% of ivermectin as a starting point following a report by Tenywa et al. (Tenywa et al., 2017) that found 0.01% of ivermectin in 10% w/v sugar solution was enough to kill more than 90% of malaria vector (Anopheles arabiensis) within 48 hours. To make a 0.01% ivermectin solution, 1ml of 1% injectable ivermectin (Ivomec®) was diluted in 100ml of 10% w/v sugar solution. The procedure was repeated to obtain other ivermectin concentrations: 0.001%, 0.0025%, 0.005%, 0.015%, 0.02%, 0.025%, 0.03%, 0.04%, 0.05% and 0.06%. The solutions were dyed with a food colouring dye (Carmoisin) at 0.5% v/v concentration for easy visualization of sugar fed mosquitoes.
Experiments were done using two rounds of five cages (30cm × 30cm × 30cm) placed inside the insectary. Each ivermectin concentration was poured into a 30mL plastic container to 2/3 of its capacity. These containers are regularly used for delivering glucose to mosquitoes reared in the insectary. A Whatman paper (filter paper) was rolled up like a tube and then dipped into each container containing the sugar solution with ivermectin concentration. The sugar solutions were randomly placed into each of the 30cm × 30cm × 30cm cages without blinding the researchers. One cage with 10% w/v sugar solution without ivermectin served as a control for the experiment. The sugar solution progressed up the filter paper where the mosquitoes had access to it. 40, 3-6 days old, blood naïve and starved for 6-8 hours mosquitoes were introduced into each cage and allowed to feed on the soaked filter paper dipped into the container containing ivermectin in 10% sugar solution. For an entire experiment, a total of 2200 mosquitoes were used as per (Tenywa et al., 2017). Mosquito mortality was recorded after 4-, 8-, 24-and 48-hours post-introduction of the treatments. For each ivermectin concentration, five experiments (replicates) were performed while changing the position of the cages after each experimental replicate to avoid bias due to cage positioning. Similar experimental set up and procedures as described above were performed for male Ae. aegypti where eight ivermectin doses were tested: 0.005%, 0.01%, 0.015%, 0.02%, 0.025%, 0.03%, 0.04% and 0.05%. In this round of the experiment, 1800 mosquitoes were used.
Semi field experiment Efficacy of ATSB on population density and survival of Ae. aegypti in a controlled environment. The efficacy of ATSB was determined in a semi field system at Ifakara Health Institute's mosquito city in Ifakara. The semi field system (biodome) is 19m × 29m long; sitting on a concrete slab and surrounded with a water moat which prevents mosquito predators from entering the biodome. Its floor is of mud and the walls are made of net which allows ambient conditions inside the biodome to be similar to the outside environment. The biodome has two compartments separated by a 4m × 29m corridor and is roofed with polyvinyl sheets (Ferguson et al., 2008). A hut mimicking rural houses in terms of doors, windows, eaves, structure and size is installed in each of the two biodome compartments. The huts' are made of red brick walls and grass roofs. To reduce the temperature inside the biodome that could lead to mosquito desiccation the biodome was modified by adding a thatch layer beneath the polyvinyl roof. The entry doors were sealed with double nets and fixed with a zip in order to prevent mosquitoes from escaping the system when a person entered or exited the biodome.
The two compartments of the biodome were labelled as A and B. Five control-ATSB (ATSB without ivermectin) and ATSBs were made following Tenywa et al., (Tenywa et al., 2017) and deployed into each of the compartments. 100 Ae. aegypti mosquitoes aged 2 -5 days, blood naïve and starved for 6-8 hours were released into each compartment at 6:00 am every day consecutively for 30 days, simulating every day emerging mosquitoes in the wild population. In compartment A, the five control-ATSBs were deployed in all 30 days of the experimentation while in compartment B, the 30 days of experimentation were done in three phases each of 10 days. During phase I (day 0 to 10 of experimentation) and phase III (day 21 to 30 experimentation) control-ATSB were placed in the compartment, while during phase II (day 11 to 20 of experimentation) the control-ATSB were replaced with ATSBs containing 0.03% ivermectin. To maintain quality of the control-ATSB and ATSB, both were replaced with a new one after every five days. 24 hours after every mosquito release, one Biogent (BG) sentinel trap was deployed into each compartment for three hours from 6:30am to 9:30am to collect mosquitoes for monitoring the Ae. aegypti population density. Mosquitoes collected from each compartment were killed with ethanol by spraying them, counted and recorded.
During phase II, human landing catches (HLC) were done in compartment A and B for 20 minutes from 6:00am to 6:20am before deployment of the BG sentinel trap at 6:30am to assess the impact of ATSB on Ae. aegypti survival. One volunteer wearing shorts, closed shoes, a long sleeve shirt and hat, sat inside each of the semi field compartment and conducted a human land catch to collect mosquitoes landing on his shins ( Figure 1). To reduce bias due to differences in volunteer attraction to the mosquitoes, the volunteers were swapped every day between the two compartments. All collected mosquitoes from each compartment were kept in paper cups labelled with the collection date and compartment/treatment. The collected mosquitoes were transferred to the insectary and given 10% w/v sugar solution that is used for regular colony maintenance. Sugar was delivered by soaking a small ball of cotton wool into 10% w/v sugar solution and put on top of the netting covering the paper cup lid. Daily mortality was recorded for 30 days. After the 30 days, the mosquitoes that remained alive were sprayed with ethanol and discarded.

Field experiment Preparation of attractive targeted sugar bait (ATSB) with and without a carbon dioxide source.
ATSB were prepared following procedures described previously (Tenywa et al., 2017) ( Figure 2A). Three test bait stations were prepared as follows: 1) control bait consisted of plain water dyed with 0.5% v/v red food colouring dye; 2) test-ATSB consisted of 10% w/v of sugar solution and ivermectin (final concentration 0.03% w/v); and 3) test-ASTB+CO 2 consisted of the same bait as test-ASTB

Figure 1. Illustration of attractive targeted sugar bait (test-ATSB or control-ATSB) stations deployed in a semi field.
The control-ATSB and test-ATSB were deployed into either compartments A or B for 30 days with replacement of the control-ATSB by test-ATSB at day 10-20 in compartment B to determine its efficacy against Ae. aegypti population density in a controlled environment. One Bio-gent sentinel trap was deployed into each semi field compartment every day at 6:30am -9:30am for 30 days. A volunteer conducted human land catch in each compartment from day 11 to 20 of the experimentation for 20 minutes from 6:00am to 6:20 am.

Figure 2. Attractive targeted sugar bait stations (test-ATSB stations). A)
test-ATSB station without a wire mesh covering the lid; B) test-ATSB station with a glue painted wire mesh covering the lid and carbon dioxide source; C) test-ATSB station with a glue painted wire mesh covering the lid. but with addition of a carbon dioxide (CO 2 ) source. The carbon dioxide was obtained from yeast. For that purpose, a total of 250grams of sugar (sucrose) and 17.5grams of yeast were dissolved into 2.5 litres of water in a 5 litre bottle. The bottle's lid was drilled to make a small hole where a small pipe was inserted into and connected to the test-ATSB+CO 2 bait ( Figure 2B). All the baits were covered with a wire mesh that was painted with rat glue (Tack-Tick Stronghold Glue ® ) to make a sticky trap ( Figure 2C) trapping any mosquito attracted to land and feed on the bait.

Application of ATSB in a field setting.
Field experiments were conducted between November and December 2019 in Tegeta, Kinondoni district in Dar es Salaam; the largest city and economic centre of Tanzania. The city is located at 6.48'S and 39.17'E along the Indian Ocean coast with 1100mm annual rainfall. It is a hub for international travel into and out of Tanzania and experiences frequent dengue outbreaks (Vairo et al., 2016). A survey was conducted to identify Ae. aegypti breeding sites using BG sentinel traps deployed twice a day: 7:00am -11:00am and 4:00pm -7:00pm. Two vehicle garages with old used tires were identified as appropriate study sites and labelled as garage A and garage B. The garages were approximately 100 metres away from each other.
Three test-ATSB and test-ATSB+CO 2 stations prepared as described above were deployed at either garage A or garage B. The stations of the same treatment were placed at 10 metres apart. To each treatment arm (test-ATSB and test-ATSB+CO 2 ), three control baits: a bait that consisted of plain water dyed with 0.5% v/v red food colouring dye, were placed; each at three metres away from each treatment station. The treatment and control stations were left at the study sites for 24 hours. After 24 hours, the baits were checked for the presence of the mosquitoes that stuck on the bait station traps, removed and morphologically identified. In order to account for any variations due to difference in mosquito densities between the study sites, the test-ATSBs and test-ATSBs+CO2 were swapped after every 24 hours between the sites. The experiments ran for a total of 12 days. And after every three days, the baits (test-ATSBs and test-ATSBs+CO2) and control were replaced with a new one so as to maintain the quality of the baits.

Data analysis
All data obtained were analysed using STATA package (Stata corp, College Station, TX).
Ivermectin LD90 against Ae. aegypti A mean cumulative proportion mortality of mosquitoes for each ivermectin concentration was determined and compared to control (10% sugar solution) at 4, 8, 24 and 48 hours.
ATSB efficacy against Ae. Aegypti in the semi-field and field Poisson regression model was employed to compare the number of mosquitoes caught in the semi-field compartments that had ATSBs vs control-ATSBs. Mosquito ages, compartments and day were considered as covariates. For the field test, a negative binomial regression model was performed to compare the number of mosquitoes that quested sugar from test-ASTB+CO 2 and test-ASTB, and control bait consisted of plain water. The incidence rate ratio (IRR) and 95% confidence interval were obtained from the model.
ATSB efficacy against wild Ae. aegypti Attractive targeted sugar baits (ATSBs) without carbon dioxide source optimized in a semi field system did not differ in attracting wild Ae. aegypti compared to a bait with plain water (IRR = 0.7; [95%CI: 0.36-1.17] and P-value ≤ 0.15) ( Table 2). However, when a carbon dioxide source was added into the  ATSB, the incidence rate ratio of captured mosquitoes increased by approximately 4.4 folds (IRR = 6.8; [95%CI:4.11 -11.30] and P -value < 0.001) in comparison to bait with sugar and ivermectin (IRR = 1) ( Table 2).    In order for sugar baits to effectively reduce mosquito population in the field, they must release enough attractant volatiles to efficiently attract vectors and there with overcome competition from the natural sugar sources. Also, application strategy (bait stations or vegetation spraying) in the field impacts sugar bait efficiency. Vegetation spraying approach involves spraying the whole vegetations with toxic sugar solution, whereas the bait station approach presents the toxic sugar bait at one single point. In this context, bait station approach may hypothetically be regarded as less effective in attracting mosquitoes due to the fact that mosquito may take a longer time to allocate the solution. However, in spite of the hypothesised limitation, we think that bait stations minimise the possibility of targeting non-targeted organisms such as butterflies, bees, ants and others, making the strategy environmentally safer.

Limitation and safety consideration
The ivermectin concentration used in the test-ATSBs in this study is relatively high; although it is unlikely that the bait concoction would be consumed by children, we recommend that further field research done on the baits should investigate these by either placing a protective grill over them or hanging them out of reach.

Conclusions
Ivermectin-based ATSBs successfully reduce laboratory Ae. aegypti populations by approximately 95% within three days in a controlled environment (semi-field) but failed to so against wild Ae. aegypti in the field unless a carbon dioxide source was added. We recommend further research to improve bait station design, and attractant strategies such as plant-based volatiles. The abstract of this study by Tenywa et al states that this study 'intended to develop and test an ivermectin-based attractive-targeted sugar bait (ATSB) against Aedes aegypti, yet this does not reflect the results presented and it is important that the authors correct this. The results of the laboratory study demonstrate the killing effect of ivermectin fed in a sugar meal to male and female Ae. aegypti. The semi-field study suggests, despite issues with the experimental design which I discuss below, that Ae. aegypti released into a biodome in the absence (presumably, it is not mentioned) of competing sugar sources appeared to feed on sugar plus ivermectin in a prototype feeding station. The presence of devices with ivermectin over a 10-day period reduced the number (or percentage) of mosquitoes caught in daily collections relative to other time periods and the control compartment. Finally, the field experiment showed that free flying mosquitoes attempted to feed on stations inside garages and were caught on the sticky mesh lid of the device, and that a greater number of mosquitoes attempted to feed when a CO 2 source was attached to the stations. Thus, the study demonstrates toxicity of ivermectin and the attractancy of CO 2 , using a prototype bait station which still has some serious development needs. Firstly, the stations seem to need to be replaced every 5 days, which is likely not feasible in full scale deployment of an intervention. There is also a safety issue and an environmental issue to nontarget organisms to consider, with open containers of sugar water with a mesh grill kept at ground level with no means that I can see to protect children or other organisms from exposure.

Ethical approval
Whilst this is an early proof of principle that ivermectin could be used in an ATSB deployment strategy, there is very little evidence of efficacy of this prototype and certainly it does not represent a final product which could be deployed. The results of the laboratory and semi-field experiments are confounded in the Conclusion, since a 95% reduction in population was certainly not demonstrated in the semi-field environment, and the comment that the stations failed to reduce populations in the field experiment is false since the experimental design was not set up to show efficacy. It is essential that the work presented is framed correctly, to correctly reflect the results of the studies, and not overclaim the significance of the results -as it stands the conclusion is that an ATSB using ivermectin can only work with a CO 2 source, which is a claim potentially damaging to innovation in the development of ivermectin-based ATSBs. Critically, this study shows that CO 2 attracts mosquitoes to the stations, but not that it encourages feeding -stations may therefore mostly target blood feeding females who may not sugar feed, and without being confident that all of those females that are attracted feed and die you are potentially increasing the biting risk to people living nearby these stations. In reality, no attractant was included in the bait stations, and so it is inaccurate to label them ATSBs or expect them to compete with natural sugar sources and be effective in population control. On the subject of terminology, the Introduction refers to 'Attractive Targeted Sugar Baits', a name trademarked by Westham Co for their device. The generic term is 'Attractive Toxic Sugar Baits' which is more relevant here since there is nothing about the device that targets them to mosquitoes specifically.
Beyond the framing and accurate representation of the results, which I strongly believe must be corrected, I think there are issues with the experimental design of the semi-field and field experiments, as well as the analysis and visualisation of data which should also be addressed.
The major issue with the semi-field experiment is that the recapturing regime is problematic -100 mosquitoes are released into the compartments at 6am, and at 6:30am a Bioagent trap is added to carry out collections and in phase 2 this included an additional HLC between 6am and 6:20am. There seems to be to be a high likelihood that a proportion of the captured mosquitoes would be the same ones which had just been released. It is thus confusing that 100% mortality was measured in the mosquitoes recaptured from the compartment with the toxic baits -unless the captured mosquitoes were screened for sugar feeding before mortality was scored, which should be mentioned in the Methods if so. In fact, a food dye was included in the experiments according to the Methods, but no data is presented on feeding rate for any experiment in the Results. Since there is no clear out of mosquitoes each day, with 50-60% of mosquitoes recaptured according to Figure 4 (if the Y axis does represent %, but in any case 100 mosquitoes were released every day), there would be 40-50% of mosquitoes remaining in the compartments from one day to the next, and so the populations would be expected to keep growing, unless there was 40-50% mortality in the compartments each day, which does not seem to match the data in Figure 4, or unless there was perfect recapture of all mosquitoes each morning which should be stated in the Methods if so.
I also have concerns about the field experiment, which is essentially a demonstration of the bait stations as a sticky trap, with and without an attractant (CO 2 ), and not a test of efficacy of the stations as seems to be claimed. There is no way to know which of the mosquitoes that were trapped would have fed on the sugar, nor any attempt to look at the effect on the population, particularly since the treatments were regularly swapped during the experiment. Given that, it's a minor point, but with the large number and high density of stations in small sites, the risk of trapping out needs to be addressed -there isn't a clear trend for reduced catches in the meta data but in many cases a large catch was followed by a lower catch the following day, particularly in garage 2. A major issue with this experiment is the inclusion of water controls and the way the data from them is handled. The captures in the water only controls were surely affected by the presence or absence of CO 2 in the nearby stations, since CO 2 will act as an attractant over an area of a few meters. Yet in the analysis the data from all water controls from both sites seems to have been combined, without accounting for the double trapping effort this constitutes. Doing some calculations from the meta data, when only compared against water on days when a competitor station was tested then ATSB alone is significantly more attractive than water, though still not as attractive as an ATSB with CO 2 . Thus it is not accurate to show, as in Table 2, that the bait without CO 2 attracted as many mosquitoes as water alone.
In several places the meta data does not match the results or methods presented in the paperfor example the text states that in the field experiment stations were swapped every 24 hours, but it seems from the meta data this was done every 3 days after day 1. The text states that the field experiment was conducted between November and December 2019, but the meta data suggests December 26 th to January 6 th . This needs to be carefully checked and misrepresentations corrected.
There has been a mistake made in the plotting of Figure 5, which does not match the text -the graph suggests only ~15% died in the ATSB compartment and 100% in the ASB compartment, whereas the text states 100% and 7.5% mortality, respectively. There is also an issue with the meta data presented for this experiment, where the sum values are incorrect in the summaries of the replicates. The numbers reported alive and dead in each replicate differs from the total numbers reported in the summary, so that the total number dead exceeds the number alive at the previous time point.
More minor issues to be resolved: Methods: More information is needed about the mosquito colony used for the lab and semi-field experiments -origin, rearing details, quality control measures, insecticide resistance phenotype, any known resistance markers.

○
In the first lab experiment and the semi-field experiment it is not clear whether male, female or both sexes were used. In the field experiments, why were the treatment stations 10m apart, but the control stations at a distance of 3m? ○ A lot more detail is needed in the data analysis section, and there are a few questions, for example, since compartment is a proxy for treatment in the semi-field experiment, why is it included as a fixed effect? Why is mosquito age included as a covariate when the same age of mosquito were used for all experiments? I am not sure that the data from lab experiments supports the first sentence regarding males and females being equally sensitive to ivermectin, since more males than females were killed within 24 hours at the lower concentrations.

○
The dose response shown in Figure 3 is also not clear, and so as shown does not seem to support the second sentence. ○ 'This study demonstrated that ATSB deployed inside a semi-field system remarkably reduced Ae. aegypti populations…' -I am not sure the data support such a strong claim.
Certainly there was a reduction in recaptured mosquitoes in Phase 2, but without knowing whether a stable population had been established in the biodome which was then reduced by the deployment of stations the exact effect is not clear.
The difference in survival time of the mosquitoes from the two groups suggests that the majority of the mosquitoes were attracted by the ATSBs and fed on it.' -to justify this claim we would need to understand the method more clearly, for example were the recaptured mosquitoes scored for sugar feeding and mortality only measured in those that had fed? What numbers of fed v unfed mosquitoes were captured? ○ 'In the field, the ivermectin-based sugar bait stations were inefficient at attracting wild Ae. aegypti.' -as discussed above, if the analysis was repeated with the water only controls separated by treatment there is some evidence of attractancy, albeit less than with the inclusion of CO 2 . Also as discussed above, there was no attractant included in the stations so I am not sure an attractive effect would have been expected. In this paragraph it would be useful to cite the recent paper about competition between ATSBs and natural sugar

Results.
"Sugar solution containing ivermectin 0.03% (w/v) caused approximately 80% of mosquito mortality for both mosquito sexes within 24 hours and > 90% within 48 hours (Figure 3)." -Data in Figure 3A does not fully support this sentence, the female 24h mortality is around 60% not 80%.

Is the study design appropriate and is the work technically sound? Yes
Are sufficient details of methods and analysis provided to allow replication by others? Partly

If applicable, is the statistical analysis and its interpretation appropriate? Partly
Are all the source data underlying the results available to ensure full reproducibility? Partly Are the conclusions drawn adequately supported by the results? Yes

Mosquitoes
"The mosquitoes were reared at 27±5ºC and 40%-99% humidity at Ifakara Health Institute insectary in Bagamoyo and Ifakara, Tanzania" -Very high variation of temperature and humidity. Was this range just for lab tests or were they combined with data from semi-field tests as well? Try to describe it in more detail.
○ "Similar experimental setup and procedures as described above were performed for males (…)" -Was the procedure described done with females? Please clarify in the text, as in other insects (kissing bugs, for example) the males also feed on blood.
○ Semi field experiment "To maintain the quality of the control-ATSB and ATSB, both were replaced with a new one after every five days." -Did you do any previous tests evaluating the stability of ivermectin mixed with sugar? How did you establish the exchange every 5 days? It is important to describe these data to ensure that the ivermectin is not being degraded or that there is loss through evaporation, since the temperature can reach 32ºC and the humidity is below 50%. ○ Did you manage to do any preference tests, even in the laboratory, to assess the preference for control bait (ASB) and toxic bait (ATSB) in the same cage or room? It would be important data to assess the preference of mosquitoes for the bait compared to other sources of natural sugars, especially in the absence of CO 2 .

Conclusions
"We recommend further research to improve bait station design and attractant strategies such as plant-based volatiles." -In addition, to plant volatile compounds, you could think of other more specific attractant compounds for mosquitoes or even for Aedes aegypti as