Effect of bendiocarb based (Ficam® 80% WP) Indoor Residual Spraying on feeding patterns from malaria vectors in Burkina Faso, West Africa.


 Background: The fight against vector is essential in malaria prevention strategies in several endemic countries in Africa. In Burkina Faso, malaria transmission is seasonal in most parts of country, so a single round of spraying should provide effective protection against malaria, provided the insecticide remains effective over the entire malaria transmission season. The outcomes of indoor residual spraying towards curtailing malaria transmission are firstly to decrease the life span of vector mosquitoes and also to reduce the malaria vectors density.Methods : CDC light trap and early morning collections by pyrethrum spray catches were performed monthly to determine the change in malaria vector indices in sprayed (Diebougou) and unsprayed sites (Dano). The female’s malaria vectors collected by both methods were used to determine their blood feeding, biting and sporozoites rate and malaria transmission risk estimated by entomological inoculation rate.Results: Anopheles gambiae complex composed to Anopheles gambiae, Anopheles coluzzii and Anopheles arabiensis were present throughout the transmission season, but An. gambiae was the predominant species collected (P =0.0005), comprising 88% of the total collected and the most infected species. Malaria vectors densities were significantly lower in sprayed villages (n=4,303) compared with unsprayed villages (n=12,569) during post-spraying period (P = 0.0012). In addition, mean human biting rate of An. gambiae sl and An. funestus sl were significantly lower in sprayed areas compared to unsprayed areas (P<0.05). Overall, malaria vector transmission risk was significant lower in villages which received IRS (P=0.0001) whatever the malaria vectors species (An. gambiae sl and An. An. funestus sl).Conclusions: The results showed that in the sprayed area (Diebougou), vector densities, human biting rates and malaria transmission risks were very lower than unsprayed areas (Dano). The findings also showed a change in vector behavior especially within An. funestus which became more zoophagic following IRS. The indoor residual spraying could be recommanded as control tool in areas where malaria transmission occured a given period of year.

The indoor residual spraying could be recommanded as control tool in areas where malaria transmission occured a given period of year.

Background
Vector control is one of the key elements of malaria control strategies [1]. In Africa, vector control relies mainly on two effective and complementary tools: long-lasting insecticidal mosquito nets (LLINs) and indoor residual spraying (IRS) [2]. Several studies have demonstrated the effectiveness of both tools in reducing the incidence of malaria-related morbidity and mortality [3,4] in Africa [5][6][7][8][9].
However, these tools, in particular LLINs, are impregnated with insecticides from the pyrethroid family. Unfortunately, the recent evolution and the expansion of resistance to this class of insecticide in West Africa in Anopheles gambiae s.s. is a major problem for sustainability in malaria prevention in Africa [10,11]. For this reason, the search for alternative tools using a non-pyrethroid insecticide has become a necessity. [12]. Indeed, since 2006, the World Health Organization (WHO) encouraged a scale-up of IRS for vector-borne disease control, using one of several classes of insecticide that are suitable for IRS [13]. In 2008, of 108 malaria-endemic burdened countries, 44 reported the use of IRS [14]. The outcomes of indoor residual spraying towards curtailing malaria transmission are firstly to decrease the life span of vector mosquitoes and also to reduce the malaria vectors density [15].
There are several insecticide formulations currently prequalified by WHO for IRS; namely organophosphates, carbamates, pyrethroids and neonicotinoids. Moreover, the effectiveness of the IRS depends on many factors, such as the remanence of the insecticide in formulations for the IRS, the feeding behavior of malaria vectors resting inside houses and susceptibility of vectors local populations to insecticide used for the IRS [16]. Residual efficacy of IRS formulations and local vector susceptibility to the insecticide used for IRS were discussed in a separate manuscript about bendiocarb efficacy on walls. The monitoring of behavioral responses of mosquitoes to insecticides is very important to the understanding of how chemicals work in the control of disease transmission [17]. Indeed, in northern Nigeria [18], IRS led to a high decrease in the total vector population and showed also a reduction in the incidence of malaria among children, the plasmodic index and fever, and an apparent effect on mortality of 1-4-year-old children. In Kenya, IRS with fenitrothion in Kisumu town [19] and the LLINs use in the south coast of Kenya [20], showed a decrease in An. funestus s.l and An. gambiae s.l. populations due to the IRS, while the high bed net coverage was followed by a much reduced human biting rate and a diminishing role of An. gambiae s.s. in malaria transmission.
In Burkina Faso, malaria was the most common cause of outpatient consultations (41.3%), hospitalizations (21.4%) and death (16.4%) in 2018 [21]. Furthermore, the resistance to at least one insecticide had been identified in 64 malaria-endemic countries according to WHO [22],, including Burkina Faso [23][24][25][26][27]. Pyrethroid resistance is particularly widespread in Burkina Faso, with high frequency of voltage gated sodium channel mutations reported as long ago as 2006 [28]. However, susceptibility tests performed in 2010-2012 with insecticides belonging to carbamates class (such as 0.1% bendiocarb) have shown a high mortality rate of the local population of malaria vectors and low allelic frequencies of gene ace-1 R [29] throughout the country including the southwest (Diebougou).
This insecticide molecule was chosen to be applied on walls during pilot study as described in the previous article. Indeed, the preliminary results from this study have shown a short persistence and a vectors resistance to this insecticide. A decision from the international community has been taken now to prioritize national and regional elimination with a long-term ultimate goal of malaria eradication [30], with the need to understand the biological implications of IRS is become paramount.
It is important to understand its impact on the behavior of malaria vectors in contact with walls treated with insecticide. In this context, the present study aims to evaluate the effect of a large scale IRS using a non pyrethroid insecticide of the carbamate family (bendiocarb) on the malaria vectors behavior in the intervention areas (Diebougou district sprayed) compared to control area used (Dano district unsprayed).  1). Villages chosen in Diebougou and Dano districts were selected to be representative of the different settings of the areas (peripheral, central, sub-urban, presence of water source, etc) and had the same type of walls such as "banco" (a mixture of mud and water) and cement. Goizper IK Vector Control compression sprayers with flat jet nozzles in order to spray onto walls and non-metal ceilings of eligible structures in IRS-targeted areas. In addition, spray operators, team leaders and spray supervisors were trained prior to spray operations. The active ingredient selected for spraying onto the walls in Diebougou was bendiocarb dosed at 80% in the wettable powder form (Ficam® 80% WP). It was applied at 400 mg of bendiocarb active ingredient/m 2 on walls of houses as recommended by the World Health Organization Pesticide Evaluation Scheme (WHOPES) [31,32].

Implementation of intervention
Prior, spray operation staff informed residents to stay out of the structure for at least two hours after IRS application.

Mosquito sampling and identification
Following IRS, the researche institute (Institut de Recherche en Sciences de la Santé, IRSS) monitored the efficacy of IRS on entomological indicators of malaria transmission. Monthly collections were conducted to determine mosquito species composition, biting rates and indoor resting densities. The malaria vector populations dynamic was monitored in each of the eight selected villages (four sprayed and four unsprayed controls) using indoor and outdoor CDC light trap collection (CDC LT) and pyrethrum spray catch (PSC). Two months of baseline collections were conducted, before IRS was performed, in June and a week before IRS application in July 2012 in Dano and Diebougou. The relatively short baseline period of data collection was due to the long dry season in the study area from November until May, during which Anopheles densities and malaria transmission is low.
Subsequent monthly entomology surveys were conducted until December 2012.
Indoor CDC LT were installed about 1.5 m above the ground next to an occupied (bait) untreated mosquito net. Unbaited light traps were also hung outdoors at the same height, approximately 10-20m away from the houses. CDC light trap was the preferred trapping method due to concerns about potential disease transmission risks during human landing catch. Moreover, several studies [33,34,35] showed a comparability of CDC light trap catch size compared with HLC for different Anopheles species. In each village, CDC light trap collections were conducted in indoors and outdoors between 20:00 pm and 06:00 am for a total of four nights per month. Four randomly selected houses were sampled each month, resulting in 16 trap-nights indoors and 16 trap-nights outdoors per month, per site. In addition, mosquitoes resting indoor were collected by pyrethrum spray catches (PSC) in four randomly selected houses in each village. PSC was performed by laying white sheets on the floor and furniture before spraying a commercial aerosol consisting of 0.64% Pyrethrum EC and 0.75% chlorpyrifos ethyl. PSC collections were performed from 06:00 to 09:00 am. The mosquitoes sampling was carried out in the same bedrooms and at the same frequency during the intervention period.

Laboratory processing
All anophelines were separate and assigned to species based on morphological characters using standard identification keys [36]. Legs of each An. gambiae sl female collected using CDC light trap and PSC methods were tested by PCR for molecular identification of species [37]. The heads and thoraces of host-seeking females were tested by enzyme-linked immunosorbent assay-circumsporozoite protein (ELISA-CSP) for Plasmodium falciparum detection using the protocol of Wirtz et al., [38]. The blood meals source from freshly fed females collected using early morning collections (PSC method) were used to assess An. gambiae sl host preference. Therefore a random selection of 30 specimens per month and per district were tested by a direct ELISA bloodmeal source detection [39] using anti-host immunoglobulin G (IgG) conjugated against human, bovine, pig, donkey and sheep blood. Unfortunately, parity rates could not be assessed because females died while in traps and were too dry for dissection. All the mosquito samples collected were stored individually in numbered vials with desiccant.

Data analyzis Measured parameters
The mean human biting rate (HBR) was calculated for each species collected by CDC LT by dividing the total number of captured specimens by the total person-nights for the collection period. The mean indoor resting density (IRD) was defined as the total number of mosquitoes (per species) collected by PSC divided by the total number of rooms sampled. The circumsporozoite (CSP) rate was calculated as the proportion of mosquitoes infected with P. falciparum sporozoites. The malaria vectors anthropophilic rate was calculated as the proportion of female mosquitoes with human blood out of the total tested for blood-meal source. The entomological inoculation rate (EIR) was calculated by multiplying the HBR indoor/outdoor and the CSP rate.
All the measured parameters were computed and analysed using the free software GraphPad 5.0 version. Data were compared with the Pearson chi 2 or Fisher exact tests and odds ratio were calculated to determine the impact of IRS in study sites. All calculations were expressed with the statistical significant threshold set at P ≤ 0.05.

Ethical considerations
Ethical approval for this study was granted by the Ethical Committee of Health Ministry in Burkina Faso. The mosquito collectors gave prior informed consent and prior they subjectd yellow fever vaccination. An partnership with health facilities close to sites was also obtained for the free antimalarial treatment of mosquito collectors who could contract from malaria.

Results
Density and species composition of malaria vectors From June to December 2012, a total of 26.276 mosquitoes (13.555 anopheline and 12.721 other culicines) were collected using both CDC light trap (9158 mosquitoes) and PSC collection methods (17.118 mosquitoes). In addition, 9.404 mosquitoes were collected in Diebougou (sprayed area) between June and December (Table 1) whose 3040 mosquitoes collected in baseline (June-July 2012) and 6.364 mosquitoes during post-spraying period (Augut-December) compared to unsprayed area with 16.872 collected mosquitoes whose 4303 mosquitoes at baseline and 12569 mosquitoes in postspraying period (P = 0.0012). a significant difference was observed between the total number collected in sprayed and unsprayed areas. According to species composition, An. gambiae sl (69.34%) and An. funestus ss (24.16%) were the most predominant Anopheline species collected in sprayed area (Diebougou) compared to 45% An. gambiae sl, 19% An. funestus ss, and 36% other culicines (Culex sp, Aedes sp., Anopheles sp etc) (Additional file 1: Figure S1). Their proportion were significantly reduced between sprayed and unsprayed areas (P = 0.039). In addition, there was a greater number of culicids collected in sprayed areas compared to unsprayed areas certainly due to impact of IRS. Overall, the total number of collected mosquitoes in sprayed areas (6364 mosquitoes) compared to unsprayed area (12569 mosquitoes) was significant (P = 0.001) after spraying period.
During the post IRS study period, indoor resting densities of malaria vectors were significantly lower in sprayed villages (n = 1,798) compared with unsprayed villages (n = 8,607) with P = 0.0051 (Table 1)

Malaria vectors monthly biting and resting behaviour following IRS Baseline data
In Dano (unsprayed), in June 2012, indoor human biting rate of An. gambiae sl was estimated at 4.6 bites per person per night by CDC light trap collection and 13.6 bites/person/night in July 2012 (Fig. 3A). However, in Diebougou An. gambiae sl human biting rates of. by indoor CDC light trap were found to be lower at 2 and 5 b/p/n indoors in June and July respectively. The An. funestus indoor human biting rate was less than 1 b/p/n (Fig. 4A). A similar trend was recorded for indoor resting densities, with Dano having approximately double the catch size of Diebougou (Fig. 5A&B). The highest resting densities by indoor PSC collection with a mean value in July reaching 49 An. gambiae sl per house per night in Dano (Fig. 5A). The catch size was generally low in outdoor CDC light trap collections in both sites (Fig. 3B & 4B).
Post-spraying data A summary of mean biting rates is presented in Fig. 3 for An. gambiae sl and Fig. 4 for An. funestus ss. In addition, the number of mosquitoes collected by month and by site is summarized in Table 2&3.
In the period post-IRS (August to December) the mean indoor biting rate per person per night (b/p/n) was significantly highest in the unsprayed sites (mean = 6.55 bites per person per night from August to December) compared to sprayed sites (mean = 3.18 bites per person per night) in An. gambiae sl (P = 0.015). furthermore, the peak from indoor biting density of An. gambiae sl occurred in August with about 40 bites per person per night in Dano (unsprayed) and decreased progressively to December, when it was less than 5 b/p/n towards the end of the rainy season (Fig. 3A). The similar pattern was observed in the intervention area but with less than 15 b/p/n of An. gambiae sl. The  (Fig. 3B). But, the exposure to mosquito bites outdoors was slightly, but more increased in Diebougou (sprayed area) after treatment compared to Dano, the unsprayed area but the difference was not significant (P > 0.05).
Similar results were observed in An. gambiae s.l. and An. funestus ss biting rates (Fig. 4) in PSC collection (Fig. 5B) with a mean biting rate of 2 b/p/n in Diebougou compared with 4 b/p/n in Dano during the post-spraying period August-December (Odds ratio (indoors CDC LT) = 0.28 with 95% CI: [0.11-0.35] and P = 0.035 (Table 3).

Plasmodium falciparum circumsporozoite and entomological inoculation rates (EIR) from An. gambiae sl and An. funestus ss
The results of CSP-ELISA assays and entomological inoculation rate of An. gambiae s.l. and An.
funestus ss are presented in Tables 2 and 3 respectively. Overall, 2051 An. gambiae sl and 1072 An. funestus ss specimens were screened for the circumsporozoite protein from June to December 2012 in the two areas. The sporozoites and entomological inoculation rate (EIR) were calculated by grouping the indoor and outdoor collections of An. gambiae sl and An. funestus ss due to low number sampled sporozoites rates detection (Table S1 in Additional file 2; Table S2in Additional file 3). So, during the post-IRS period (August-December), the mean sporozoites rate differed between unsprayed and sprayed areas for CDC LT method. The average sporozoites rates were significantly different (more than 2-fold) between the unsprayed areas (average sporozoite rate = 15.82%; 95% CI: [8.94-23.49]) and the sprayed areas (average sporozoites rate = 6.05%; 95% CI: [3.509-12.59]) (t = 2.475; df = 9 with P = 0.022) ( Table 2). The highest sporozoites rates were observed in Dano in August (average sporozoite rate = 13.6%; 95% CI: [9.68-17.33]) and September reaching an average of 14.2%. The similar trends were also observed in An. funestus ss sporozoites rate (Table 3) but in lowest proportions (average sporozoite rate = 2.14%; with 95% CI: [0. 25-3.15] in unsprayed areas and average sporozoites rate = 1% with 95% CI: [0. 33-1.8] in sprayed area) with a significant difference (P = 0.035).
The major contributor to the EIR, both in the control and intervention areas, was An. gambiae s.l.
(70%). The indoor EIR reached 134 infective bites/person during the five-month post-IRS in the unsprayed area Dano. IRS appears to have reduced the EIR four-fold in the sprayed area (31 infective bites /person) after spraying with P = 0.0001. An. funestus ss contributed also to the transmission in the two areas, with the similar results (EIR reduced 4-fold in sprayed area) compared to An. gambiae sl after spraying (mean EIR in Dano = 13 bi/p/n vs mean EIR in Diebougou = 3 bi/p/n with P = 0.003).

An. gambiae sl and An. funestus ss blood meal sources
The results presented in Figs. 8 and 9 included data from indoor and outdoor CDC LT and indoor PSC collections, for An. gambiae sl and An. funestus ss from the two areas. Irrespective of the sampling month, the proportion of An. gambiae sl. blood-fed on human was highest, reaching more than 80% of the total of 335 females analysed, both in the sprayed and unsprayed areas. No female was recorded blood-fed only on animals. For An. funestus ss the feeding patterns were quite different, especially in the sprayed area (Diebougou) where females showed a large range of hosts. Out of 132 An. funestus ss females analysed for their blood-fed origin, about 20-40% had taken a mixed bloodmeal (human, bovine and oat). A potential effect of the IRS on An. funestus ss was that the proportion of human blood meals decreased being replaced by animal and mixed blood meals.

Discussion
The study showed that two species An. gambiae s.l. and An. funestus ss were predominant vectors of malaria transmission in study areas [28] collected using CDC LT and pyrethrum spray catches. the results also indicated that the An. gambiae sl entomological inoculation rate was 4-fold lower in sprayed area compared to the unsprayed area, after the implementation of the IRS primarily due to a lower indoor biting rate and a significant decrease of malaria vectors sporozoite rates. However, the indoor resting density of vectors declined in the sprayed houses following IRS, but increased slightly in October, probably due to the relatively short residual duration of bendiocarb indicated in a separate manuscript. This drastic drop could be also due to the lethal effect of bendiocarb on the anophelines resistant to pyrethroids [40]. The biting rates observed outdoors were slightly higher in intervention areas compared to control areas and may be an early sign of biting behavior change. In conclusion IRS did not reduce the endophily behaviour from that of the baseline but had significantly reduced the density of mosquitoes resting indoors in sprayed area compared to the unsprayed area.
In addition, EIR might have been impacted by the observed outdoor biting behaviour, and the reduced residual efficacy of the insecticide after September. The overall indoor biting rate of An. funestus ss was twofold greater in the unsprayed area. Indeed, the susceptibility status, taxonomy, and the role of An. funestus ss in malaria transmission was well documented in similar areas at west (Lena) and southwestern Burkina Faso (Soumousso) [44].
The use of vector control tools and behaviors of the host would be the main factors that modify the behavior of human blood feeding observed on An. gambiae sl. Indeed, recent studies [41,42] showed that the long-term indoor application of residual insecticides contributes towards an increased tendency for outdoor feeding among malaria vector populations. The treatment had a great positive impact within An. funestus decreasing the human host-seeking in interventions sites compared to control sites. It is probably due to behavior change that this species is opting to go outside to seek a bloodmeal. It is important to highlight the exophagic host seeking activities and the exophilic behaviour that resulted in the search for blood in animals and mixed meals analyzed which more developed by An. funestus ss in the intervention area than in the control area where this vector remains mainly endophagic and endophilic. Indeed, after treatment there is less of An. funestus ss collected in intervention area compared to control area. Moreover, the exophagic host seeking recorded in intervention area was more pronounced than that obtained with An. gambiae s.l. Gillies and De Meillion [36] stated "…Funestus shows a closer adaptation to human dwellings than any other African anopheline. In many areas it spends the greater part of its adult life in houses, which has made it one of the most vulnerable of species to attack with residual insecticides". This statement is clearly valid in this case. Such a clear response to indoor insecticides makes the emergence of insecticide resistance in this species all the more likely [43]. Moreover, the results of impact of IRS on malaria transmission by An. gambiae sl indicated that the transmission was lower compared to the control area, after the implementation of the IRS, where the biting rate of Anopheles dropped drastically. This drastic drop is due to the lethal effect of bendiocarb on the anophelines resistant to pyrethroids [40] even though such efficacy did not last more than three months (discussed in a separate manuscript). The biting rates observed in outdoors were higher in intervention areas compared to control areas assuming a less pronounced behavior change of vectors.
In conclusion IRS did not reduce the endophily behavior from that of the baseline though it had significantly reduced the density of mosquitoes resting indoors compared to the control area. In addition, the findings have also shown An. gambiae sl were particularly anthropophilic in the two areas with few cases of mixed blood meals and no pure animal blood meals identified. This feeding pattern was the inverse for An. funestus ss in the intervention area where more females showed a large plasticity of the host range (zoo-anthropophilic). The results have shown that An. gambiae s.s.
S-form was the major malaria vector species biting in the southwestern region in Burkina Faso. This corroborates previous reports [28 -29] of the anopheline distribution in Burkina Faso, which explained the abundance of the An. gambiae species by the ecological characteristics of this region.

Conclusion
The pilot study of IRS with bendiocarb appeared to have a significant impact on malaria transmission in the sprayed areas, as measured by EIR. Indeed, the results illustrated that IRS was strong enough to reduce mosquito abundance, sporozoite rate, and EIR in pyrethroids resistance areas. However, the baseline period indicated intrinsic differences in biting rates between Dano and Diebougou before IRS.
The findings also showed a change in vector behavior, with An. funestus becoming more zoophagic after IRS. Furthermore, the residual efficacy of IRS did not last more than three months. In areas of

Ethics approval and consent to participate
Ethical approval for this study was granted by the Ethical Committee of the Ministry of Health in Burkina Faso. The mosquito collectors gave prior informed consent and they were vaccinated against yellow fever. They were also subjected to regular medical check-ups with preventive treatments of malaria.

Consent for publication
Not applicable.

Availability of data and materials
Data supporting the conclusions of this article have been included within the article. Raw data will be made available upon request to the corresponding author.

Competing interests
The authors declare that they have no competing interests.       Proportion of An. gambiae sl and An. funestus ss blood-fed on humans, animals or mixed from sprayed area (Diebougou) and unsprayed area (Dano).

Supplementary Files
This is a list of supplementary files associated with this preprint. Click to download.
Additional file 1.pdf Additional file 2.pdf Additional file 3.pdf