Combined use of long-lasting insecticidal nets and Bacillus thuringiensis israelensis larviciding, a promising integrated approach against malaria transmission in northern Côte d'Ivoire

Background The recent reduction in malaria burden in Côte d’Ivoire is largely attributable to the use of long-lasting insecticidal nets (LLINs). However, this progress is threatened by insecticide resistance and behavioral changes in Anopheles gambiae sensu lato (s.l.) populations and residual malaria transmission, and complementary tools are required. Thus, this study aimed to assess the efficacy of the combined use of LLINs and Bacillus thuringiensis israelensis (Bti), in comparison with LLINs. Methods This study was conducted in the health district of Korhogo, northern Côte d'Ivoire, within two study arms (LLIN + Bti arm and LLIN-only arm) from March 2019 to February 2020. In the LLIN + Bti arm, Anopheles larval habitats were treated every fortnight with Bti in addition to the use of LLINs. Mosquito larvae and adults were sampled and identified morphologically to genus and species using standard methods. The members of the An. gambiae complex were determined using a polymerase chain reaction technique. Plasmodium infection in An. gambiae s.l. and malaria incidence in local people was also assessed. Results Overall, Anopheles spp. larval density was lower in the LLIN + Bti arm 0.61 [95% CI 0.41–0.81] larva/dip (l/dip) compared with the LLIN-only arm 3.97 [95% CI 3.56–4.38] l/dip (RR = 6.50; 95% CI 5.81–7.29; P < 0.001). The overall biting rate of An. gambiae s.l. was 0.59 [95% CI 0.43–0.75] biting/person/night in the LLIN + Bti arm against 2.97 [95% CI 2.02–3.93] biting/person/night in LLIN-only arm (P < 0.001). Anopheles gambiae s.l. was predominantly identified as An. gambiae sensu stricto (s.s.) (95.1%, n = 293), followed by Anopheles coluzzii (4.9%; n = 15). The human-blood index was 80.5% (n = 389) in study area. EIR was 1.36 infected bites/person/year (ib/p/y) in the LLIN + Bti arm against 47.71 ib/p/y in the LLIN-only arm. Malaria incidence dramatically declined from 291.8‰ (n = 765) to 111.4‰ (n = 292) in LLIN + Bti arm (P < 0.001). Conclusions The combined use of LLINs with Bti significantly reduced the incidence of malaria. The LLINs and Bti duo could be a promising integrated approach for effective vector control of An. gambiae for elimination of malaria. Supplementary Information The online version contains supplementary material available at 10.1186/s12936-024-04953-8.


Background
The burden of malaria remains a significant concern in sub-Saharan Africa, despite the progress made in controlling this disease during the two last decades [1].The World Health Organization (WHO) has reported recently that there were 249 million cases of malaria worldwide and an estimated 608,000 malaria deaths in 2023 [2].The WHO African region accounted for 95% of global malaria cases and 96% of deaths attributable to malaria, with the most severely affected being pregnant women and children under 5 years old [2,3].
Long-lasting insecticidal nets (LLINs) and indoor residual spraying (IRS) have played a key role in reducing the malaria burden in Africa [4].The scaling-up of these malaria vector control tools led to a 37% drop in the number of malaria cases and a 60% reduction in the mortality rate between 2000 and 2015 [5].However, the post-2015 trend is worryingly stagnant or even increasing and the level of malaria deaths remains unacceptably high, particularly in sub-Saharan Africa [3].Several studies have underlined that the emergence and spread of resistance of the major Anopheles vectors of malaria to the insecticides used in public health is an obstacle to the future effectiveness of LLINs and IRS [6][7][8].In addition, the change in the behaviour of the vectors to biting outdoors and earlier at night is responsible for residual transmission of malaria and is of increasing concern [9,10].The limitations of LLINs and IRS in controlling the vectors responsible for residual transmission are a major weakness in current efforts to eliminate malaria [11].Moreover, the persistence of malaria is attributed to climatic conditions and human activities that contribute to the creation of larval habitats [12].
Larval source management (LSM) is a vector control approach based on targeting mosquito breeding sites with the objective of reducing both the number of breeding sites and the number of mosquito larvae and pupae they contain [13].LSM has been recommended by several studies as a complementary integrated strategy for malaria vector control [14,15].Indeed, the efficacy of LSM offers the dual advantage of targeting malaria vector species that bite both indoors and outdoors [4].Additionally, vector control through LSM based on a biolarvicide like Bacillus thuringiensis israelensis (Bti) can increase the panel of malaria control tools.Historically, LSM has played a key role in the successful control of malaria in the USA, Brazil, Egypt, Algeria, Libya, Morocco, Tunisia and Zambia [16][17][18].
Although, it has played a primary role in integrated pest management in several countries that have eliminated malaria, LSM has not generally been integrated into malaria vector control policy and practices in Africa, being only used in the vector control programmes of a few sub-Saharan countries [14][15][16][17][18][19].One reason behind this is that breeding sites are generally considered to be too numerous and difficult to locate, thus making LSM application very expensive to implement [4-11, 13, 14].As a result, the WHO has recommended for decades that the resources mobilized for malaria vector control should be focused on LLINs and IRS [20,21].Only in 2012 did the WHO recommend the integration of LSM, especially Bti intervention, as a complement to LLINs and IRS in certain settings in sub-Saharan Africa [20].Since this WHO recommendation, several pilot studies have been conducted on the feasibility, efficacy and cost of applying biological larvicides in sub-Saharan Africa, and these have demonstrated the efficacy of LSM in reducing Anopheles mosquito densities and malaria transmission [22][23][24].
Côte d'Ivoire is among the top 15 countries in the world where more than three quarters of the global malaria burden occurs [25].The malaria prevalence in Côte d'Ivoire represents 3.0% of the world malaria burden with an estimated incidence and cases ranging from 300‰ to over 500 ‰ inhabitants [25].Malaria transmission is perennial in the northern savannah region of the country despite the long dry season from November to May [26].Malaria transmission in this region is linked to the presence of a high proportion of asymptomatic carriers of Plasmodium falciparum [27].In the region, the most abundant malaria vectors are Anopheles gambiae sensu lato (s.l.) populations.Local An. gambiae s.l. are predominantly by Anopheles gambiae sensu stricto (s.s.), which is strongly resistant to insecticides, thus leading to a high risk of residual malaria transmission [26].Because of local vector resistance to insecticides, the use of LLINs can may have a limited impact in reducing malaria transmission which, therefore, remains of great concern in this region.Pilot studies based on use of either Bti or LLINs have shown effectiveness in lowering mosquito vector densities in the northern region of Côte d'Ivoire.However, no studies have previously assessed the impact of repeated application of Bti combined with the use of LLINs on malaria transmission and on malaria incidence in this region.Therefore, the present study was designed to evaluate the impact of the combined use of LLINs and Bti on malaria transmission by comparing a LLIN + Bti arm with a LLIN-only arm in four villages in the northern region of Côte d'Ivoire.The hypothesis was that a Bti-based LSM would add value by further reducing malaria mosquito densities when implemented in addition to LLINs compared the LLINs only being used.Such an integrated approach, targeting both Anopheles immatures with Bti and adults with LLINs, may be crucial in reducing malaria transmission in an area of high malaria endemicity as the villages of northern Côte d'Ivoire.The outcomes of this study could thus help inform a decision on the possible integration of LSM into the national malaria vector control programmes (NMCPs) of endemic sub-Saharan countries.
The malaria prevalence was higher 82.0%(2038 cases) in the target population in the Napié health centre (data before Bti intervention).In all four villages, householders used only PermaNet ® 2.0 LLINs distributed by the NMCP of Côte d'Ivoire in 2017, with a coverage rate of > 80% [25][26][27][28]30].The villages belonged to the Korhogo region, which is a sentinel site of the NMCP of Côte d'Ivoire and are accessible all year round.Each of the four villages includes at least 100 households, with comparable populations and several cases of malaria were recorded there each year as indicated in the health registers, a working document of the Ministry of Health (MoH) of Côte d'Ivoire).Malaria is mainly caused by P. falciparum transmitted to people by An. gambiae s.l. with transmission also by Anopheles funestus and Anopheles nili in the region [28].The local An.gambiae complex consists predominantly of An. gambiae s.s. with high the frequencies of the kdr mutation (frequency range: 90.70-100%) and moderate frequencies of the ace-1 allele (frequency range: 55.56-95%) [29].
The average annual rainfall and temperature ranged from 1200 to 1400 mm and from 21 to 35 °C, respectively, and the relative humidity (RH) was estimated at 58%.The climate of this study area is Sudanian type, with a 6-month dry season (November to April) and a 6-month rainy season (May to October).The area is experiencing some of the impacts of climate change, such as the disappearance of vegetation cover and a longer dry season marked by the drying up of water bodies (lowlands, rice paddies, ponds, puddles) that can serve as larval habitats for Anopheles mosquitoes [26].
The study was conducted in the LLIN + Bti arm represented by the villages of Kakologo and Nambatiourkaha and the LLIN-only arm represented by the villages of Kolékaha and Lofinékaha.In all of these villages, people were using only PermaNet ® 2.0 LLINs during this study period.

Study design
The effectiveness of use of LLIN (PermaNet 2.0) in combination with Bti against Anopheles mosquitoes and malaria transmission were evaluated through a randomized controlled trial (RCT) with two study arms, LLIN + Bti arm (treatment arm) and LLIN-only arm (control arm).The LLIN + Bti arm was represented by Kakologo and Nambatiourkaha, while Kolékaha and Lofinékaha was designed as the LLIN-only arm.In all four villages, the local populations were using PermaNet ® 2.0 LLINs received from the NMCP of Côte d'Ivoire in 2017.The use condition of PermaNet ® 2.0 was assumed similar between the villages, because they have received the net in the same way.In the LLIN + Bti arm, Anopheles mosquito larval habitats were treated with Bti every fortnight in addition to the LLINs already being used by the populations.The larval habitats were treated within the villages and within a 2-km radius from the centre of each village as recommended by the WHO and followed by NMCP of Côte d'Ivoire [31].In contrast, no larviciding treatments with Bti were conducted in the LLIN-only arm during the study period.

Larviciding treatments
The water-dispersible granule formulation of Bti (Vectobac WG, 37.4% w/w; lot no.88-916-PG; 3000 International Toxic Units IUT/mg; Valent BioScience Corp, USA) was used at a dose of 0.5 mg/L.A knapsack sprayer containing 16 L with a fiberglass lance supplied with handle and adjustable nozzle with a flow rate of 52 mL per second (3.1 L/min) was used.For preparing a sprayer containing 10 L of water, the quantity of Bti diluted in the suspension was 0.5 mg/L × 10 L = 5 mg.For example, for treating an area with a water volume estimated at 10 L in water with a sprayer containing 10 L, the quantity of Bti to dilute was 0.5 mg/L × 20 L = 10 mg.The 10 mg of Bti were measured using electronical balance in the field.The suspension was prepared mixing this Bti quantity in the graduated pail of 10 L using a spatula.This dose was chosen after a pilot field trial on the efficacy of Bti on different larval instars of Anopheles spp.and Culex spp.under natural conditions in a separate but similar area as the current study area [32].The amount of suspension of larvicide applied per breeding site and the time duration of application were calculated taking into account the estimated volume of water in the breeding sites [33].Bti was applied using calibrated hand-held sprayers.The sprayers were calibrated and tested in separate exercises and different areas to ensure that the desired amounts of Bti were delivered.
To find the best time to treat the larval breeding sites, the team determined a spray window.A spray window is the period during which the applied product provides optimum efficacy: in this study, it lasted from 12 h to 2 weeks, depending on the persistence of Bti.The period, from 7 a.m. to 6 p.m., seems necessary to enable larvae present in a breeding site to ingest the Bti.Thus, periods of heavy rain were avoided, and when rain meant that spraying had to be stopped treatment was resumed the next day if the weather was favourable.The dates to spray, as well as the exact day and time depended on the weather conditions observed.To calibrate the knapsack sprayer according to the required Bti treatment application rate, each technician was trained to visually check and position the sprayer nozzle and maintain the pressure.Calibration was completed by checking that the correct amount of Bti treatment was applied evenly per unit area.The larval habitats were treated every fortnight.Larviciding activities were done with the support of four experienced and well-trained technicians.The larviciding activities and participants were supervised by a highly experienced supervisor.The larviciding treatments were started during the dry season, in March 2019.Indeed, the dry season was indicated in a previous study as a the most appropriate time period for the larviciding intervention due to the stability of the breeding sites and their reduced numbers [27].In anticipation, the control of the larvae during the dry season could prevent the pullulation of the mosquitoes during the rainy season.Two (02) kilograms of Bti valued at $99.29 enabled all sites to be covered in the study arm treated.In the LLIN + Bti arm, larviciding interventions lasted for a whole year, from March 2019 to February 2020.In total, 22 larviciding treatment events occurred in the LLIN + Bti arm.

Side effect assessment
The potential side effects (e.g., itching, dizziness or nose running) were monitored among the Bti biolarvicide sprayers and the occupants of the households involved in the LIN + Bti arm through individual interviews.

Net use rate
Household surveys were conducted among 400 households, made up of 200 households per study arm to assess the percentage of LLINs used in the population.The household surveys were based on quantitative approach using a questionnaire.The rate of use of LLINs was stratified in three age categories: < 5, 5-15 and > 15 years.The questionnaire was addressed and explained in the Sénoufo local language to the heads of the households or other adults over 18 years.

Net use rate =
Person number having slept on net * 100 Total person having slept in hours on the eve of survey The minimum size of households to be surveyed was calculated using the formula described by Vaughan and Morrow [34].
n is the sample size, e is the margin of error, t is the margin coefficient deduced from the confidence rate, p is the proportion of the elements of the parent population that have a given property.Each element of this fraction has a conventional value, thus (t) = 1.96; e = 0.05.The minimum size of households in this condition to survey was 384.

Entomological surveys Larval collections
Before initiating the current trial, the different types of Anopheles mosquito larval habitats were identified, sampled, described, georeferenced and labelled in both LLIN + Bti and LLIN arms.The size of the breeding sites was measured using a tape measure.Then, the density of mosquito larvae was assessed every month for 12 months at 30 breeding sites randomly selected per village, totaling 60 breeding sites per study arm.There were 12 larval sampling events corresponding to 22 treatment of Bti in each study area arm.The choice of these 30 breeding sites in each in each village was intended to track an equitable number of larval collection points by village and study arm to minimize the biases.A dipper of 60 ml was used to collect the larvae by the dipping method [35].Use of a small dipper, unlike the standard WHO dipper (350 mL), was necessary due to the specificity of some breeding sites which were very small and shallow.Totals of 5, 10 or 20 dips were taken from the breeding sites of perimeters < 1 m, 1-10 m and > 10 m, respectively.The collected larvae were identified morphologically to genus (e.g., Anopheles spp., Culex spp.and Aedes spp.) directly in the field [36].Collected larvae were divided into two categories according to development stage: early instar (stage 1 and stage 2) and late instar (stage 3 and stage 4) larvae [37].Larvae were counted per genus and per stage of development.After counting, mosquito larvae were reintroduced into their specific breeding sites that were refilled to their initial volume with the original water and topped up with rainwater.
The breeding sites were considered positive when at least one larva or pupa of any mosquito genus was present.The larval density was determined by dividing the number of larvae per genus by the number of dips.

Adult collection
Adult mosquitoes were collected every two months from ten randomly selected households in each village on two consecutive days per survey.During three consecutive days, a total of 20 households were sampled per study arm throughout the study.Mosquitoes were captured using standard window trap (WT) and pyrethrum spray catch (PSC) methods [38,39].First, all houses in each village were numbered.Then, four houses were selected randomly as the adult mosquito collection sites in each village.In each randomly selected house, mosquitoes were collected in the main bedroom.The selected bedrooms had both doors and windows and were occupied by humans on the previous night.The bedrooms were kept closed before to start and during the mosquito collections to prevent mosquitoes from flying out of the room.A WT was installed on each window of each bedroom selected as mosquito sampling point.The next day, mosquitoes that had exited the bedrooms into the WTs were collected early in the morning between 06:00 a.m. and 08:00 a.m.Mosquitoes were collected from the WTs using a mouth aspirator and stored in disposable cups covered with a piece of untreated mosquito net.After WT-based collections, mosquitoes resting inside the same sampled bedrooms were immediately captured using pyrethroid-based PSC.After placing a white sheet on the floor of the bedroom, the doors and windows were closed and the insecticide (active ingredients: 0.25% transfluthrin + 0.20% permethrin) was sprayed.Approximately 10-15 min after spraying, the cover was removed from the sprayed bedroom and mosquitoes that had fallen onto the white sheet were collected with forceps and stored in petri dishes containing water-soaked cotton.The number of people who had stayed overnight in the bedroom sampled was also recorded.The collected mosquitoes were rapidly transferred to the field laboratory for further processing.

Laboratory procedures
In the laboratory, all the collected mosquitoes were identified morphologically to genus and species [36].The ovaries of An. gambiae s.l.females were dissected in a drop of distilled water on slides, using a binocular dissecting microscope [35].The parity status was scored to separate parous females from nulliparous females according to morphological aspects of the ovarial tracheoles and determine their parity rate and physiological age [35].
The anthropophilic index was determined by detection of the origin of bloodmeals of freshly blood-fed An. gambiae s.l.females among human, livestock (bovine, sheep, goat) and chicken hosts by enzyme-linked immuno-sorbent assay (ELISA) bloodmeal technique [40].Entomological inoculation rate (EIR) was calculated with human blood fed An. gambiae s.l.females estimation [41] In addition, An. gambiae s.l.infection with Plasmodium malaria parasites was determined by analysing the head and thorax of parous females by using the ELISA-circumsporozoite antigen detection (ELISA-CSP) method [40].Finally, members of the An.gambiae complex were identified by analysing their legs, wings and abdomen by polymerase chain reaction (PCR) technique [34].
The formulas of key entomological outcomes measures used are [38]: EIR is the Entomological infection rate.

Malaria clinical data collection
Malaria clinical data were obtained from the clinical consultation registers of the Napiélédougou health centre that covered all four villages (i.e., Kakologo, Kolékaha, Lofinékaha and Nambatiourkaha) included in the present study.The review of registers was focused on the records from March 2018 to February 2019 and March 2019 to February 2020.The clinical data from March 2018 to February 2019 have constituted the base data or data before Bti intervention and these from March 2019 to February 2020 were the data after Bti intervention.The clinical information, age, and village of each patient in both LLIN + Bti and LLIN study arms were collected in the health registers.Information such as the village origin, the age, and the diagnosed pathology of each patient was recorded.The case considered in this study as malaria was confirmed by rapid diagnostic tests (RDTs) and/or malaria microscopy test with prescription of artemisininbased combination therapy (ACT) by a health agents.The malaria cases were subdivided into three age categories (i.e., < 5, 5-15 and > 15 years).The malaria incidence rate per 1000 inhabitants per year was estimated by dividing the rate of the malaria prevalence per 1000 inhabitants by the population size of the village.
The formula of malaria incidence is: in between both study arms using the LLIN + Bti arm as the baseline.Effect sizes were expressed as odds ratio and 95% confidence interval (95% CI).The rate ratio (RR) of Poisson test was used to compare the proportions and the malaria incidence before and after Bti intervention in each study arm.The significance level used was 5%.

Ethical considerations
The study protocol was approved by the National Ethics Committee for Research of Ministry of Health and public Hygiene in Côte d'Ivoire (N/Ref: 001//MSHP/CNESVSkp) and the health district and administrative authorities of the region of Korhogo.Signed informed consent was obtained from the household survey participants and the owners and/or residents before collecting mosquito larvae and adults.Household and clinical data was anonymized and kept confidential, with access restricted to the designated investigators only.

Effects of Bti on mosquito larvae Mosquito larval habitats and composition
A total of 1198 breeding sites were visited.Among these breeding sites inspected in the entire study area 52.5% (n = 629) were in the LLIN + Bti arm and 47

Anopheles spp. larval density
The overall mean larval density of Anopheles spp. was 0.61 [95% CI 0.41-0.81]larvae per dipper (l/dip) in the LLIN + Bti arm and 3.97 [95% CI 3.56-4.38]l/dip in LLIN-only arm during this trial (Additional file 1: Fig. S1).The mean density of Anopheles spp. was 6.5-fold high in LLIN-only arm than in LLIN + Bti arm (RR = 6.49; 95% CI 5.80-7.27;P < 0.001).During the treatment period, no one Anopheles spp.larvae was collected in the LLIN + Bti arm from January corresponding to twentieth Bti treatment.The significant reduction in larval density in LLIN + Bti arm was observed in both the early and late instars.
Early instar Before the Bti treatment starting (in March), the mean density of early instars of Anopheles was estimated at 1.28 [95% CI 0.22-2.35]l/dip in the LLIN + Bti arm and 1.37 [95% CI 0.36-2.36]l/dip in the LLIN-only arms (Fig. 2A).After the Bti treatment was applied, the mean density of early instar of Anopheles generally decreased gradually from to 0.90 [95% CI 0.19-1.61] to 0.10 [95% CI − 0.03-0.18]l/dip in the LLIN + Bti arm.In the LLIN + Bti arm, the density of Anopheles larvae in the early instars has remained low.In the LLIN-only arm, a fluctuation of Anopheles spp.Larvae in the early instars was observed, with a mean density from 0.
Late instar Before the Bti treatment, the mean density of late instar Culex larvae was 0.97 [95% CI 0.09-1.85]and 1.60 [95% CI − 0.16-3.37]l/dip in LLIN + Bti and LLIN arms, respectively (Fig. 3B).Following the initiation of Bti treatments, the mean density of late instar Culex spp. in the LLIN + Bti arm decreased progressively and became lower than in LLIN-only arm where densities remained very high.The mean density of late instar Culex larvae was 0.12 [95% CI 0.07-0.15]l/dip in LLIN + Bti arm and 1.36 [95% CI 1.11-1.61]l/dip in the LLIN-only arm.The mean density of late instar Culex larvae was significatively higher in the LLIN-only arm than in the LLIN + Bti arm (RR = 11.19;95% CI 8.83-14.43;P < 0.001).

Effects of Bti on adult mosquitoes and malaria transmission Species composition
A total of 3456 adult mosquitoes were collected in the study area.Mosquitoes belonged to five genera (Anopheles, Culex, Mansonia, Aedes and Eretmapodites) and 17 species (Table 1).Among the malaria vectors, An. gambiae s.l. was the most abundant species, with a proportion of 74.9% (n = 2587), followed by An. funestus (2.5%, n = 86), and An.nili (0.7%, n = 24).The abundance of An. gambiae s.l. was lower in the LLIN + Bti arm (10.9%, n = 375) than in the LLIN-only arm (64%, n = 2212).No An.nili individuals were collected in the LLIN-only arm.However, An. gambiae and An.funestus were present in both the LLIN + Bti and LLIN-only arms.

Culicid nuisance
In

Malaria transmission
Anopheles gambiae s.l. was the most abundant malaria vector in the study area.The biting rate of An. gambiae females was 0.64 [95% CI 0.27-1.00]b/p/n in the LLIN + Bti arm and 0.74 [95% CI 0.30-1.17] in the LLIN-only arm at the beginning of study (Fig. 6).During the Bti intervention, the highest biting rate activities were observed in September month corresponding to twelfth Bti treatment, with a peak of 1.A total of 646 An. gambiae were dissected.Overall, the percentage of local An.gambiae parous was generally > 70% throughout the study except in July in the LLIN-only arms (Additional file 3: Fig. S3).However, the mean parity rate was 74.5% (n = 481) in study area.In LLIN + Bti arm, the parity rate has remained high and superior at 80% to exception September month where it reduced to 77.5%.Whereas a fluctuation was observed in parity rate mean in LLIN-only arm, with lowest parity rate mean estimated 64.5%.

Net use rate
A total of 400 households with a population of 2435 inhabitants were surveyed.The average density was 6.1 persons per household.The LLINs ownership rate in households was 85% (n = 340) compared with 15% (n = 60) which did not own an LLIN (RR = 5.67 [95% CI 4.29-7.59],P < 0.001) (Additional file 5: Table S5).The LLIN use rate was 40

Impact of Bti intervention on clinical malaria incidence
A total of 2484 clinical cases were recorded in the Napié health centre for the period from March 2018 to February 2020.Clinical malaria prevalence in the general population represented 82.0% (n = 2038) of all clinical cases of pathologies.Local malaria incidence per year was 479.8‰ and 297.5‰ before and after Bti treatment in the present study area (Table 2).

Perceived side effects
There were no negative side effects reported among the sprayers and the people of the of the villages treated with Bti biolarvicide either during or after applications.

Discussion
Recent progress in the reduction of malaria burden, which has largely been attributable to deployment of LLINs and IRS, is now threatened by Anopheles vector resistance to insecticides and behavioural changes, residual transmission, and low use of LLINs [2].Thus, the current study aimed at producing evidence on the possible benefits of combining LLINs and Bti-based larviciding in reducing local An.gambiae s.l.densities and malaria transmission and clinical incidence in comparison with LLINs alone in the rural savannah villages of northern Côte d'Ivoire.The results of the current study showed that repeated applications of Bti biolarvicide to Anopheles mosquito larval habitats in the presence of LLINs reduced significantly the larval and adult densities of An. gambiae and the transmission and number of clinical cases of malaria in the LLIN + Bti arm compared with LLIN-only arm.These outcomes demonstrated the added value of integrating biolarvicide-based LSM programmes into the mass distribution of LLINs to combat residual malaria transmission in rural sub-Saharan African settings.
The current field study was conducted in high malaria transmission areas of northern Côte d'Ivoire where Anopheles mosquitoes and other several key arboviral and filarial disease vectors (Cx.quinquefasciatus and Ae.aegypti) co-occur.Indeed, while only Anopheles mosquitoes were targeted in this study, the collected Culicidae fauna comprised 17 species, with a predominance of the malaria primary vector An.gambiae and presence of secondary vectors An. funestus and An.nili.The high species diversity and high abundance of mosquitoes in the study area could be attributed the presence of various types of larval habitats and to a limited impact of LLINs distributed through the national campaigns of mass LLIN distribution led by the NMCP of Côte d'Ivoire.The LLINs use rate among the target population was 76.9% with a net use rate very low in each study arm.This low rate of net use observed could be explained by the fact that this study was carried out during the second year after the distribution.The wear of the fabrics could lead to the abandonment of certain nets.It is also likely that part of the population would not like to use the LLINs because of the heat or that the LLINs are used for other purposes.Although access to LLINs is made possible and facilitated thanks to the joint efforts of governments and partners, the sustainable mass use of this control tool remains a challenge.For a better protection of the population against the infective bites of the vectors, the LSM in complement to LLINs should be an option of choice in the decision-makers.
In this study, local mosquito immatures (larvae and pupae) were collected from aquatic habitats, including rice paddies, reflows of showers, earthenware vessels, river edges, pools, village pumps, puddles, animal hoof prints, swamps, jars, ponds, and watering wells.The predominance of the highly anthropophilic An. gambiae s.l., composed mainly of An. gambiae s.s.(95%) co-existing with An. coluzzii, could be explained by the existence of rice fields and puddles but also frequent human bloodfeeding opportunities as only < 40% of local households were using LLINs.Such a low use rate of LLINs (< 80% LLIN universal coverage recommended by WHO) is common in rural sub-Saharan villages and could be due to limited ownership and access to LLINs, and not using nets due to the heat.The abundance of human-biting and parous An. gambiae populations combined with a low use of LLIN may explain the high transmission of P. falciparum and high numbers of clinical malaria cases found in the present study areas.Therefore, these areas In the present study, after the Bti larviciding of larval habitats was initiated in the LLIN + Bti arm, the density of Anopheles spp.and Culex spp.immatures (i.e., early instar larvae, late instar larvae and pupae) gradually decreased in the LLIN + Bti-intervention arm, leading to significantly lower mosquito larval and adult densities in the LLIN + Bti arm compared to the LLIN-only arm.Similar observations have been reported from previous studies that showed the effectiveness of a Bti larviciding intervention on mosquito immatures in Kenya [22] and Burkina Faso [23], and other African settings [23].Early studies have demonstrated the effectiveness of Bti on An. gambiae and Cx.quinquefasciatus larvae in rice growing areas of Côte d'Ivoire [31] and Tanzania [43].Bti is a biological or a naturally occurring bacterium found in soils and containing effective against the immature forms of insects, including the mosquito larvae.The Bti is ingested by larvae in the form of crystals released from the spores which are transformed into protoxin molecules.These molecules, under the action of intestinal protease in the target organism, release cytolytic (Cyt1Aa) protoxin which cross the peripheral membrane for fixing on over specific receiver situated on the intestinal epithelium surface.The accumulation of Cyt1Aa in the epithelium cell entrains firstly the alimentary canal paralysis and then the intestinal cell lysis of the mosquito larvae [44].Bti has been shown to have no toxicity to people and is approved for use for pest control in organic farming operations.There is no documented mosquito larval resistance to Bti as a biolarvicide to date [44].Bti has been found to control effectively pyrethroid resistant Anopheles mosquitoes.Indeed, a variety of δ-endotoxins confers on Bti an insecticidal potential [45].Moreover, cytolytic toxins present play an important role in the toxicity of Bti against dipteran larvae by increasing the toxicity and reducing the possibility of the appearance of resistance [46].During the dry seasons a low density of Anopheles spp. was seen in both the LLIN + Bti and the LLIN only arms during the study period, likely due to changes in biotic and abiotic factors (sunlight, mud, dilution, crystals, rainfall, flushing), as observed in Burkina Faso, Côte d'Ivoire and Keya [27,32,33].These seasonal variations are not likely to have affected negatively or biased the comparison of LLIN + Bti arm with LLIN arm as both study arms were subjected to the same environmental conditions across the seasons.Therefore, it could be advantageous to start larvicide applications during the dry season to reduce the exponential growth in mosquito numbers following the rains.In addition, the use of Bti during the dry season has the effect of reducing the amount of Bti required because the volume of water in the breeding sites is lower and the number of breeding sites during the dry season are fewer compared with the rainy season.Therefore, beginning Bti intervention during the dry season in the current study might have greatly improved its efficacy against Anopheles and Culex mosquito larvae.
The high impact of Bti applications in controlling Anopheles mosquito larvae observed in the present study significantly affected An. gambiae adult population biting rates, resulting in lower malaria transmission and clinical incidence in the LLIN + Bti compared to the LLINonly arm.Indeed, before initiating the Bti intervention, in the villages where LLIN was previously the only vector control measure, the number of human-biting An. gambiae females and malaria cases were high.Once the Bti intervention was added to the use of LLINs, both EIR and the number of malaria cases dropped markedly in the LLIN + Bti treatment arm area.This significant reduction may relate to Bti applications which controlled Anopheles larvae, preventing them from reaching the adult stage and, thus, reducing Plasmodium transmission to local people.The parity rate of An. gambiae remained high in both study arms but was higher in the LLIN + Bti arm compared with LLIN-only arm, during the Bti applications.This may be explained by the lower larval density in the LLIN + Bti arm, reducing adult emergence and leading to a higher percentage of parous females of An. gambiae in this study arm compared to the LIIN-only arm.Indeed, the Bti lethal effect on the larvae, especially on the late instar, might have inhibited the emergence of An. gambiae female neonates in the LLIN + Bti arm, which continued to occur in the LLIN-only arm [47].Parous females may also have possibly invaded the LLIN + Bti arm from neighbouring areas were Bti was not applied, as observed in Burkina Faso [47].Conversely, the adult population of An. gambiae could be regenerated by the emergence of female neonate in LLIN-only arm, in which Bti was not applied.The lowest malaria prevalence in the Bti-treated villages was consistent with the low vector density and the comparatively low EIR.Additionally, EIR obtained in the current field trial showed that people have received fewer infective bites in the Btitreated villages than in the LLIN-only villages.However, the annual EIR recorded in the Bti-treated villages was slightly high as an annual EIR above 1 ib/p/y could be enough to sustain Plasmodium transmission [48].Nevertheless, the lower vector density and the strong decline in malaria incidence in the LLIN + Bti arm could be due to supplementary protection created locally by the Bti intervention.Indeed, this positive effect of Bti in protecting people against malaria was more pronounced in children under five who are more vulnerable to malaria infections.
Although LLIN + Bti showed overall high effectiveness against An.gambiae s.l. and malaria transmission in the present study, additional investigations are needed to address methodological limitations for better understanding the epidemiology of this disease in the target the villages.This study was conducted in villages with mean nets usage rates under 40%, only.Applying Bti in villages with low nets use rates (e.g., low and middle nets use rates) would likely help to strengthen them effectiveness and more beneficial for effective malaria control programs.Plasmodium transmission decreased significantly in the villages where Bti was applied with an EIR was still high (> 1 ib/p/y) in LLIN + Bti arm thus allowing possible residual malaria transmission in this area.Assessing the mosquito dispersal and possible invasion of villages within the LLIN + Bti trial arm with adult An. gambiae mosquitoes from neighbouring untreated breeding sites or areas may help to determine the buffer zone around treated villages to prevent recolonization [49].This high rate of An. gambiae females parous in the LLIN + Bti could being justified emergence inhibition of neonate females.Moreover, assessment of Plasmodium infection among the whole populations through parasitological surveys in the villages in addition to the clinical cases among patients in the hospitals may allow to better understand the impact of the LLIN + Bti combination of malaria prevalence.Indeed, previous studies have reported that the number of Plasmodium-infected, asymptomatic carriers was high [27], or people suffering from malaria would not always go to health centre in villages.The use of Bti should be considered as an additional vector control intervention in combination with newer LLINs products (e.g., PermaNet 3.0, PermaNet Dual, Interceptor G2).Finally, large-scale trials of the combined use of Bti and LLINs to assess community acceptability and operational effectiveness against malaria vector control are needed to further validate this approach.
In summary, in the current community trial, the repeated application of Bti biolarvicide to Anopheles mosquito breeding sites in addition to LLINs in a highly endemic area was found to be highly effective in reducing Plasmodium transmission as well as malaria morbidity, Thus, integrated Bti and LLINs intervention could be seen as a promising approach to enhancing malaria vector control in Côte d'Ivoire [25].Moreover, no adverse events and complains were reported among the sprayers and the inhabitants of villages treated with the Bti biolarvicide during and after the applications, and this might improve community acceptability and adherence for combination of Bti with LLINs.The use of water as solvent can reduce the environmental impact and costs of Bti treatment [50].
The use of malaria clinical health data collected from local health registers of a working document of health centres covering the study area could be a limitation of this study.A parasitological study in the target population in the study area could have generated more reliable data on malaria incidence.The absence of vector abundance data before the Bti intervention could be a limitation of this study, as this data could have allowed us to examine the effect of the Bti intervention on change in the relative abundance of different species of mosquitoes.The fact that this study was done in the context of low net use and pyrethroid only nets is a potential limitation as it is possible that with effective control of mosquitoes through high coverage with effective nets, the impact of larviciding would be much lower.Another study on Bti including a high coverage with effective nets could be confirmed the results of this study.It is also a limitation that only 4 villages were included.Lastly, a study on the cost of LSM intervention could be essential to better guide decision-makers.

Conclusion
The current community study conducted in the savannah northern region of Côte d'Ivoire demonstrated that the repeated applications of the Bti biolarvicide to natural breeding sites of mosquitoes, in addition to the use of LLINs in households, resulted in an effective control of Anopheles larvae.This led to a significant reduction in An. gambiae density and malaria transmission and clinical incidence among the target population in the LLIN + Bti arm compared to in the LLIN-only arm.These outcomes suggested that, while the use of LLINs against vectors is indispensable as a malaria control intervention, it may require other complementary vector control tools, including larviciding, to maximize effectiveness, especially where LLIN use rates are low and where some of the malaria vectors may bite early or outdoors.Therefore, integrating a Bti-based LSM programme should be considered as a complementary measure to LLIN mass distribution to help drive malaria elimination efforts in Côte d'Ivoire and other malaria endemic settings in sub-Saharan Africa.In addition, Bti applications in the presence of LLINs had a positive effect in reducing both Anopheles and Culex mosquitoes, thus reducing malaria transmission and mosquito nuisance biting, and this may be beneficial to the optimal coverage of LLINs, as reduction of mosquito nuisance biting is key to the acceptance and adoption of vector tools.Ultimately, Bti biolarvicide applications in association with the distribution if LLINs appears to be a promising tool targeting the major vector An.gambiae larvae for the control and the elimination of malaria, even in areas with low use or coverage of LLINs.

Fig. 2
Fig. 2 Variation in the average density of larvae of Anopheles spp. of early instar (A) and of late instar (B) in the study arms, in Napié area in northern Côte d'Ivoire, from March 2019 to February 2020.LLIN: long-lasting insecticidal nets; Bti: Bacillus thuringiensis israelensis; Trt: treatment

Fig. 3 Fig. 4
Fig. 3 Variation in the average density of larvae of Culex spp. of early instar (A) and early instar (B) in the study arms, in Napié area in northern Côte d'Ivoire from March 2019 to February 2020.LLIN long-lasting insecticidal nets, Bti Bacillus thuringiensis israelensis, Trt treatment

Table 2
Impact of Bti intervention on malaria incidence from March 2018 to February 2020 in the target population in Napié area LLIN, long-lasting insecticidal nets; Bti, Bacillus thuringiensis var.israelensis; MI, Malaria incidence; ‰, per 1000 inhabitants; %, percentage n: Total malaria cases in study area; RR, Rate ration; 95%CI, Confidence interval at 95% Same letter denoted no difference between the malaria incidence before and after Bti intervention P > 0.05 Different letter denoted Difference between malaria incidence before and after Bti intervention P < 0.05 Bti biolarvicide to LLINs to control An. gambiae and malaria transmission in rural villages of sub-Saharan Africa.