Experimental huts trial of the efficacy of pyrethroids/piperonyl butoxide (PBO) net treatments for controlling multi-resistant populations of Anopheles funestus s.s. in Kpomè, Southern Benin

Background: Insecticides resistance in Anopheles mosquitoes limits Long-Lasting Insecticidal Nets (LLIN) used for malaria control in Africa, especially Benin. This study aimed to evaluate the bio-efficacy of current LLINs in an area where An. funestus s.l. and An. gambiae have developed multi-resistance to insecticides, and to assess in experimental huts the performance of a mixed combination of pyrethroids and piperonyl butoxide (PBO) treated nets on these resistant mosquitoes. Methods: The study was conducted at Kpomè, Southern Benin. The bio-efficacy of LLINs against An. funestus and An. gambiae was assessed using the World Health Organization (WHO) cone and tunnel tests. A released/recapture experiment following WHO procedures was conducted to compare the efficacy of conventional LLINs treated with pyrethroids only and LLINs with combinations of pyrethroids and PBO. Prior to huts trials, we confirmed the level of insecticide and PBO residues in tested nets using high performance liquid chromatography (HPLC). Results: Conventional LLINs (Type 2 and Type 4) have the lowest effect against local multi-resistant An. funestus s.s. and An. coluzzii populations from Kpomè. Conversely, when LLINs containing mixtures of pyrethroids and PBO (Type 1 and Type 3) were introduced in trial huts, we recorded a greater effect against the two mosquito populations (P < 0.0001). Tunnel test with An. funestus s.s. revealed mortalities of over 80% with this new generation of LLINs (Type 1 and Type 3),while conventional LLINs produced 65.53 ± 8.33% mortalities for Type 2 and 71.25 ±7.92% mortalities for Type 4. Similarly, mortalities ranging from 77 to 87% were recorded with the local populations of An. coluzzii. Conclusion: This study suggests the reduced efficacy of conventional LLINs (Pyrethroids alone) currently distributed in Benin communities where Anopheles populations have developed multi-insecticide resistance. The new generation nets (pyrethroids+PBO) proved to be more effective on multi-resistant populations of mosquitoes.

mortalities ranging from 77 to 87% were recorded with the local populations of .

An. coluzzii
This study suggests the reduced efficacy of conventional LLINs Conclusion: (Pyrethroids alone) currently distributed in Benin communities where populations have developed multi-insecticide resistance. The new Anopheles generation nets (pyrethroids+PBO) proved to be more effective on multi-resistant populations of mosquitoes.

Introduction
Malaria is responsible for about 438,000 deaths with an estimated 214 million disease cases annually 1 . Malaria vector control tools have been encouraging worldwide, resulting in a decreased morbidity and mortality as of 2016 compared to the 2000 2 . Unfortunately, as this disease has reduced globally, it has been a different case in Africa, where malaria is still a serious challenge 2 . Long lasting insecticide-treated nets (LLINs) are major components of malaria control tools, and they have helped to combat malaria disease when in good conditions and properly used 2 . LLINs are effective, simple to use, easy to deliver to rural communities, and cost-effective when used in highly endemic malaria areas 3 .
In Benin, malaria control is hugely dependent on LLINs and indoor residual spraying (IRS) 4,5 . In October 2014, there was a country-wide distribution campaign of mosquito nets to ensure universal coverage, with the free distribution of 6,077,272 LLINs to 2,199,522 households surveyed 6 . After this exercise, LLINs utilization by children under five rose from 70% in 2012, to 73% in 2014 6 . However, the emergence and spread of resistant malaria vectors to insecticidal components used for treating these nets have threatened the earlier progress made with this malaria vector control tool 7-9 . Resistance to insecticides ofone of the main malaria vectors, An. funestus against control tools has since become a serious challenge facing the quest for malaria elimination in Africa. Reported cases of resistance are available in countries such as Cameroon 10 , Uganda 11 , Mozambique 12 , Malawi 13,14 , Ghana 15 , Nigeria 16 and Benin 17,18 . There are also multiple mechanisms that are driving observed resistance in this mosquito population, although over-expression of detoxification genes remains the main driving force of insecticide resistance in this Anopheles species 15,19 . Another observation is that resistance mechanisms are known to differ from one mosquito population to another suggesting the contribution of geographical differences in resistance profiling 15,20 . This revelation is of serious concern because it is becoming a significant threat to existing malaria control tools. Recently, a study by Agossa et al. 9 in the northern part of Benin, showed that the efficacy of existing malaria vector control tools treated with pyrethroid have decreased in wild An. gambiae s.l. populations. Considering the fact that An. funestus and An. gambiae have developed resistance to almost all classes of insecticides across Benin 17,21,22 , it might follow a similar trend as the above study. Indeed, there is a serious quest for alternative insecticides since pyrethroids are becoming less effective with recorded reports of resistance in malaria vectors 23 . Pyrethroids are very safe, acceptable and suitable for LLINs, but degrade very fast, especially when exposed to sunlight, which can be avoided if nets are well preserved 24 . A different insecticide resistance management approach combining a chemical synergist, piperonyl butoxide (PBO), with pyrethroids on net fibres could be a promising way to fight insecticide resistance. PBO, a synergist capable of inhibiting the action of oxidase enzymes, has potential to combat the growing problem of oxidase based pyrethroid resistance in mosquito vectors species. Two types of long lasting nets treated with permethrin+PBO and deltamethrin+ PBO are the new generation of LLINs for improved resistance management 25,26 . These new generation nets have shown their efficacy on some resistant populations of Anopheles in experimental hut trials 25,27 . In hut trials, the new generation of LLINs increases mortality and inhibits blood feeding against pyrethroid-resistant An. gambiae in some Africa regions 23,25,26,28 . In Nigeria, the efficacy of LLINs treated with deltamethrin +PBO was highly effective on resistant An. gambiae compared with standard treated nets with no PBO 29 . Also, in Southern Africa (Mozambique), this combination proved to be more effective against resistant An. funestus and An. gambiae 12,20 .
Due to the widespread of insecticide resistance in most populations of An. funestus from South to the North of Benin 17,18,21 , it is important to assess the efficacy of currently used LLINs (conventional LLINs) and also conduct in experimental huts a comparative assessment of the performance of the new generation of treated LLINs (Pyrethroids+PBO) against conventional LLINs currently used by communities in areas of Benin where the main malaria vectors An. gambiae and An. funestus have developed multiple resistance to insecticides.

Study area
The assessment was conducted in the rural locality of Kpomè (6° 23' N, 2° 13'E) located in South Benin, approximately 81 km from Cotonou. The study area has a sub-tropical climate, receives 1,100 mm of mean rainfall annually and has a mean monthly temperature between 27 and 31°C. The rainfall pattern in this area is similar to other southern localities of Benin, with two rainy seasons and two dry seasons. The constant presence of water bodies in this locality favors the development of An. funestus and other mosquito species 18 . Previous studies carried out in Kpomè showed that An. funestus s.s. is mainly predominant during the dry season and transitional periods, and exhibited high resistance to permethrin and deltamethrin with mortalities rates (World Health Organization (WHO) susceptibility tests) of 13% and 46.5% respectively 17 . P450s are the main family of detoxification enzymes involved in observed pyrethroid resistance in the An. funestus population in this locality 17 . An. gambiae populations from this same locality have also developed multi-resistance to several insecticides families 30 . This set of available environmental and entomological data has prompted building of seven experimental huts at Kpomè for trials to identify best LLINs types for improved control of insecticide resistant populations of malaria vectors.

Collection of mosquitoes for planed experiments
Early morning collections of blood-fed, semi-gravid and fully gravid females of resistant An. funestus resting inside houses were collected using electric aspirators between 06h00 and 10h00 in June 2017 (Consent from head of household was obtained prior to collection). These mosquitoes were identified morphologically using Gillies and De Meillon 31 and Gillies and Coetzee 32 key as An. funestus were kept in small cups and immediately transported to the laboratory (Relative humidity of 70-80% and a temperature of 25-30 °C) until fully gravid (for blood-fed and semi-gravid females). Eggs were obtained from F 0 generation (Collected females from the field) using the forced egg laying method 33 and were allowed to hatch to obtain F 1 generation to be used for different experiments.

Piperonyl butoxide (PBO) synergist tests
According to the level of observed resistance against permethrin and deltamethrin, and because of pyrethroids resistant, An. funestus population has been shown to express P450s genes more than in previous studies 19,39,40 , 2-5 days old F1 mosquitoes were pre-exposed to 4% PBO paper for 1 h and immediately exposed to 0.75% permethrin and 0.05% deltamethrin for 1 h. Two controls were used during this experiment. The first control was the mosquitoes exposed to untreated papers without PBO, and the second comprised of mosquitoes exposed to paper treated with PBO only. Mortalities were recorded 24h post exposure and were later compared to the un-synergized group in order to evaluate the potential role of cytochromes P450 genes in the observed resistance.
Characteristics of long-lasting insecticidal nets used during the various assessments Five types of LLINs were used for the phase I (Cone and tunnel tests) and phase II (experimental hut) evaluations.
The Type 1 LLINs made of monofilament polyethylene (100 mesh size) fabric treated with deltamethrin at 4 g/kg±25% and piperonyl butoxide (PBO) at 25g/kg±25%, side panels made of multifilament polyester fabric with a strengthened border treated with deltamethrin at 2.1 g/kg±25%.
The Type 2 LLINs was made of multifilament polyester fabric (100 mesh size), treated with a deltamethrin only (no PBO added) at 1.4 g/kg±25%.
The Type 3 LLINs was treated with 20 g/kg of permethrin and 10 g/kg of PBO in the whole polyethylene net fibres (150 mesh size).
The Type 4 LLINs made of polyethylene fibers treated with permethrin only (no PBO) at 20 g/kg, incorporated during fibers extrusion (150 mesh size).
The Type 5 was an untreated net, a multifilament polyester (100 mesh size) fabric with neither insecticide nor PBO treatments.
All nets used had sizes of 160 cm wide, 180 cm long and 150 cm high. All types of nets treated and control nets were procured by the Liverpool School of Tropical Medicine (LSTM), UK, properly wrapped and shipped to us at IITA for our various trials.

WHO cone tests with of 5 types of nets under tests
Four cones were fixed in contact to 25 × 25 cm pieces of nets taken from the sides and top panels of LLINs (Methods in Anopheles Research, 2010). 2 to 5 days old individuals from the 3 colonies of Anopheles mosquitoes (resistant-An. funestus Kpomè, Resistant-An. gambiae Kpomè and Susceptible-Anopheles Kisumu) were exposed to nets for 3 min, after which they were transferred into recovery paper cups and provided with cotton wool soaked in a 10 % honey solution. A minimum of 50mosquitoes was tested for each net. At least three pieces per net was used for this test. Mosquito knock-down rate (kd) was recorded at 1h post-exposure period and the mortality rate was determined 24 h post-exposure. The mortality rate was corrected using Abbott's formula if needed. These tests were conducted at a room temperature and relative humidity of 25-30°C and 70-80% respectively.

WHO tunnel tests with fragments of the 5 types of nets under tests
Tunnel test was carried out with the same samples of LLINs used for cone test. Adults An. funestus, An. gambiae and Kisumu strain were also used for this test 41 . Adult mosquitoes aged between 5 to 8 days were released at 6.00 PM in the first compartment C1 of a 60-cm long tunnel made of glass divided by a transverse netting (25cm x 25 cm) insert, fitted onto a frame that slots across the tunnel. The LLINs fragments used had been pierced (1-cm diameter holes) to allow mosquitoes to pass through it into the tunnel to the compartment C2 where a guinea pig was placed for mosquitoes feeding. Each guinea pig was used only once for this study. Guinea pigs were sourced from local markets where they are sold for food consumption. At 8.00 AM of the following morning, mosquitoes were collected from both compartments and transferred into plastic cups. The mortality and feeding status (blood-fed or unfed) of each mosquitoes collected from the tunnel were recorded. Blood-feeding rate and penetration rate across the tunnel were also assessed.
Experimental hut trials of the efficacy of the 5 Net Types on insecticide resistant mosquitoes Experimental huts newly built in Kpomè are specially designed to test the efficacy of different vector control products against freely entering mosquitoes under natural but controlled conditions. This facility was used for our release and recaptures tests. Huts were typical of the West African model as recommended by WHO 41 . The 3.5 × 2 × 2 m huts were made from concrete bricks, with a corrugated iron top and a ceiling of thick polyethylene sheeting lined, and each was built on a concrete base surrounded by a water-filled moat to exclude ants. Mosquito access was through 4 window slits, constructed from pieces of iron fixed at an angle to create a funnel with a 1-cm gap, present on 3 sides of the huts. Mosquitoes had to fly upward to enter through the gaps and downwards to exit; this precluded or limited exodus through the aperture and enabled us to account for most entering mosquitoes. A veranda trap made of concrete bricks and mesh screening (2 m long × 1.5 m wide × 1.5 m high) projected from the back wall of each hut. Movement of mosquitoes between a room and the veranda was unimpeded.

Study design
The described 5 types of mosquito nets were assessed against pyrethroid resistant An. funestus and An. gambiae. The control mosquito population used was only An. gambiae Kisumu as we had neither laboratory/field susceptible An. funestus, nor field susceptible An. gambiae.

5-Type 5 Untreated Polyester net (control)
Blank assessment of hut attractiveness Prior to introducing nets into huts, we conducted preliminary experiments which showed the huts to be evenly attractive to mosquitoes. Briefly, assessment for freely-entered mosquitoes in the hut was conducted during 2 weeks and the attractiveness effects of each hut were evaluated. Adult male volunteers slept under the untreated net in the huts from 20:00 hours to 05:00 hours each night after cleaning the hut to remove any spiders and ants. To minimize biases in individual attractiveness, sleepers were rotated between huts on successive nights throughout the 2 weeks.
Blank assessment of huts lethality Still prior to assessments, an initial series of bioassays was conducted to determine the mortality of susceptible mosquitoes exposed to various surfaces in the huts to know the lethal effect of the huts. Bioassays were performed with WHO cones tests attached to the surfaces with masking tape. In each hut, surfaces tested included doors, walls, screening-mesh of veranda, ceiling and floor. Ten females of An. gambiae Kisumu strain of 2 to 5 days old were put into each cone for 30 min. After this exposure time, they were removed from the cone and put into plastic cups covered with untreated mosquito net and given access to 10% honey solution and mortalities recorded after 24h.
Release and recapture experiment A release/recapture experiment was conducted in experimental huts with resistant populations of An. funestus and An. gambiae both from Kpomè and a susceptible An. gambiae Kisumu. These 3 populations of mosquitoes were released different days into huts where the 5 described types of nets were erected. The experiment was conducted as described by WHO protocols 41 . The main trial was conducted in August 2017. The treatments were allocated randomly to five experimental huts in study site. Each net was deliberately holed with six 4cm×4cm holes to simulate a worn net. Before experimental hut evaluation, adult volunteers had been recruited among the inhabitants of the villages where experimental huts were implemented and informed consent to participate in the study was given beforehand and, chemoprophylaxis was provided during the trial. Female mosquitoes aged 5 days were released in each hut at 20:00 h in the night and monitored till morning. Early in the morning, released mosquitoes were recaptured from the hut, veranda and inside the nets and were scored as dead or alive and as fed or unfed. Live mosquitoes were kept in small cups containing sugar solution for 24 hours to assess the delayed mortality. Entomological effects of treatments were compared in-between nets and with the untreated net (control Net Type 5). Target entomological parameters monitored included: induced exiting, blood-feeding inhibition and mortality.
(i) insecticide-induced exiting, i.e. the proportion of mosquitoes found in hut verandahs relative to control huts; (ii) blood-feeding inhibition, i.e. the proportional reduction in blood-feeding relative to untreated nets; and (iii) mortality, the proportion of mosquitoes killed (immediate plus delayed).

Chemical analysis of net used in the experimental hut trial
Prior to the trial, chemical analysis were conducted on pieces of nets (pieces from holes made on nets) from the five Net Types erected in each hut. This experiment was to confirming the presence, or absence, and the concentration of pyrethroids and PBO in each net to be used in this trial. For the LLINs Type 1, the side panels and top panel were tested separately. Chemical analysis was conducted using high performance liquid chromatography (HPLC) machine (Agilent technology1260 infinity, Germany). Deltamethrin, permethrin and PBO was extracted using acetonitrile as solvent and the mixture was sonicated for 15min. Afterwards, the solution without the net was transferred into a new flask and filtered through a 0.45µm PTFE syringe filter into an HPLC vial for analysis. For HPLC analysis, standard solutions of each insecticide (Permethrin cat no. 45614, Deltamethrin cat no.45423 and PBO cat no.45626) purchased from Sigma Aldrich were prepared from stock solution in acetonitrile. Standard curve of each insecticide were drawn. The HPLC system condition was as follow: mobile phase: Acetonitrile /H2O (90:10), C18 Column, flow rate: 1ml/min, injection volume: 50µl and UV detector wavelength: 226nm.The quantities of insecticides were calculated based on the peak area and expressed in g/kg of net.

Data analysis
Data from bioassays were compared between each net using MedCalc easy-to-use online statistical software, version 18.2.1 42 , while the Fisher's exact test was used to test for significant difference of mortality rates. Significance between treatments was set at 5% level. The proportion of mosquitoes that exited early (induced exophily), the proportion that was killed within the hut (mortality) and the proportion that successfully blood-fed (blood feeding rate) were compared and analyzed using the logistic regression with treatments as fixed effects and huts, sleepers as random effects (STATA 9 Software). for permethrin and deltamethrin respectively ( Figure 1). Low mortality rates were also recorded for An. coluzzii population from Kpomè against permethrin (19.27±3.52%) and deltamethrin (60.11±7.19%). No mortality was recorded when subsets of these mosquitoes species were exposed to papers with no insecticides (Control).

Molecular speciation of
Synergist assay with PBO When permethrin and deltamethrin were combined with PBO ( Figure 2)    Blank assessment of experimental huts attractiveness A total of 603 mosquitoes were allowed to freely enter the seven experimental huts during the 12 trial nights. The mean number of mosquitoes collected in huts was high in hut N°7 (18.41) followed by hut N°2 (8.41). The mean number of mosquito per night was almost similar in huts N° 6, 5, 4 and 3. The hut N°1 showed a relatively low attractiveness (Table 4). However, similar attractiveness in terms of Anopheles mosquitoes was observed between the hut N° 1, 2 and 5. The recorded mean numbers of Anopheles mosquito in hut N° 3, 4, 6 and 7, which are similar, were higher than the others.

Blank assessment of experimental huts lethality
The cones bioassay conducted on various surfaces, such as doors, walls, screening-mesh of veranda, ceiling and floor, of each hut revealed that all the huts built in Kpomè locality had no lethal effect on susceptible Anopheles gambiae Kisumu strain. The mortality rate for all exposed mosquitoes was very low as only one mosquito died out of the total of 73 exposed in hut N°4 and N°7 (Table 5).

Release and recapture experiments
Induced exophily. When Kisumu strain was released in rooms containing treated nets, we recorded a significant movement of mosquitoes from the room to the veranda; all treated nets induced significant exophily rates ranging from 50 to 73% compared to the untreated net, where the observed exophily was 30% (P< 0.0009). A similar trend was observed with the pyrethroids resistant An. funestus s.s. population from Kpomè with exophily rates ranging from 30 to 40% with treated nets compare to 9.46% with untreated Net Type 5 ( Table 6). The induced exophily rates recorded with An. funestus s.s. was not significant in between huts containing treated nets (induced exophily rate ranging from 23 to 34%) ( Figure 4). As for the pyrethroids resistant An. coluzzii, all treated nets induced exophily rates ranging from 8% to 37% ( Figure 5).  More specifically there was blood-feeding inhibition with An. funestus s.s. populations in presence of Net Type 4 (44% blood feeding inhibition) and Net Type 2 (58% blood feeding inhibition) compared to untreated nets. Generally, higher bloodfeeding inhibition rates were provided by Net Type 1 containing deltamethrin + PBO (92% blood feeding inhibition) than Type 2 containing deltamethrin only (P<0.0001). Similarly, blood feeding inhibition rates in huts with Net Type 3 containing permethrin + PBO (100% blood feeding inhibition) was higher than those with Net Type 4 containing permethrin alone (Figure 4) ( Table 6). As for the pyrethroid resistant An. coluzzii, blood feeding was inhibited more with Net Type 1 which contains deltamethrin + PBO (76% blood feeding inhibition) than Type 2 which contains deltamethrin only (58% blood feeding inhibition). Respectively, 44% and 57% blood feeding inhibition rates were recorded in huts with Net Type 4 and Net Type 3 ( Figure 5).

LLINs Types
Induced exophily Blood feeding inhibition % Percentage 33.03% and 90.28% mortality rates were recorded in the huts containing Net Type 4 and Type 3 ( Figure 6). Consequently, the overall killing effect offer by Net Type 1 was significantly higher than Net Type 2 against resistant An. funestus s.s. (Table 6). Same thing with Net Type 3 in comparison to Net Type 4. The same trend was observed against resistant An, coluzzii, where very low overall killing effect was provided by Net Type 2 (28.49%) compared to Net Type 1 (73.65%). Mortalities rose from 10.7% to 71.8%, when resistant An. coluzzii were released in huts containing respectively Net Type 4 and Type 3. Consequently, high overall killing effect was provided by Net Type 3 against this Anopheles specie compared to Net type 4 ( Table 6). A combined pyrethroids-PBO Net Type 1 and Type 3 was found to demonstrate a greater efficacy against these resistant mosquito populations.

Discussion
This study aimed to assess the response of resistant An. funestus s.s. from Benin to pyrethroid treated nets (current LLINs) and to combined PBO + pyrethroid nets for improved control of resistant populations of malaria vectors.

Bio-efficacy of selected LLINs types
Results obtained from the response of susceptible mosquitoes (An. gambiae Kisumu) to treated nets showed that pyrethroid and pyrethroid + PBO treated nets remained effective for controlling susceptible Anopheles mosquitoes. It was also observed that the bio-efficacy of nets treated with deltamethrin only (Type 2) was significantly lower when we compared the recorded mortality rates from the cone test in the resistant populations of An. funestus s.s. and An. coluzzii and the susceptible strain Kisumu. These observations further confirm the high pyrethroid resistance observed in both malaria vectors in Kpomè like in others localities of Southern Benin 17,21 . A more recent study conducted across a South-North transect of Benin, revealed that more than 50% of An. gambiae mosquitoes are unaffected by lethal effects of the current form of Net Type 2 43 . However, in the Ivory Coast, this net was effective against An. gambiae s.s. 44 . When resistant mosquitoes were exposed to the combined deltamethrin-PBO (Net Type 1), the mortality rose from 56.67 to 95.77% and from 34.67 to 69.54% for respectively An. funestus s.s. and An. coluzzii. This finding showed the important involvement of P450s genes in observed pyrethroids resistance in this study and confirms also the results of synergist bioassays test performed with these same resistant mosquitoes as almost all individuals were dead when they were exposed to PBO and immediately after to deltamethrin.
Similarly, significantly lower mortality of An. funestus s.s. in the presence of current permethrin treated Net Type 4 was observed compared to the combined Net Type 3 (permethrin + PBO). The loss of bio-efficacy of this current Net Type 4 was also demonstrated in Malawi, Mozambique and Democratic Republic of Congo, where recorded mortality rates of An. funestus to Net Type 4 in this study were respectively 3%, 20% and 34% 12,13,20 . The study conducted in Benin in 2013 demonstrated the efficacy of combined permethrin-PBO net (Olyset plus) against resistant An. gambiae 25 . Surprisingly, only 25.83% of An. coluzzii was affected by lethal effect of this net in this study. It could probably due to the presence of other mechanisms involved in multi-resistance of An. coluzzii from Kpomè like kdr mutations 30 . This result supports the relatively low mortality (69.67%) obtained from the synergist test when we pre-exposed An. coluzzii to PBO before to permethrin. Therefore, a combined permethrin -PBO net does not provide a solution to pyrethroid resistance with An. coluzzii from Kpomè, Southern Benin.
Tunnel test performed on the all net Types used in this study confirmed the reduced bio-efficacy of only pyrethroids treated nets, showing a decrease in their effectiveness in areas of high   resistance. This observation could be related to the resistance selection pressure generated by the use and misuse of the same class of insecticides for malaria vector control in public health and for pest control in agriculture 21,45,46 . Indeed, reduced repellent effect of Net Type 2 against wild resistant Anopheles mosquitoes compared to high repellent effect against Kisumu strain could be as a result of their resistance nature. However, crossing of mosquitoes through Net Type 1was highly inhibited for each resistant population, even for susceptible strain, penetrating the compartment C2 of the tunnel containing Net Type 1. Nevertheless, deltamethrin alone used for the treated net (Type 2) continues to have moderate performance against resistant Anopheles mosquitoes in terms of reducing human and malaria vector contact and also blood feeding rate.
Crossing rates with Net Type 1 Malima et al. 52 where the recorded mortality of An. funestus was 71.6% against nets treated with permethrin only. Survival rates of these mosquitoes in the huts suggest that the protective nature of currently used net Type 4 in Benin is compromised, as previously reported 53 . These semi-field controlled experiments confirmed the results from laboratory phase I evaluations and displays faith in combined pyrethroids-PBO nets, despite the multiple resistance mechanisms present in these mosquito species 4,19,20,21,[54][55][56] . However, it is necessary to further investigate the impact of these multiple mechanisms on the efficacy of nets treated with pyrethroids only against An. funestus s.s.
A combination of the synergist PBO to pyrethroids made treated nets more efficient as PBO acted both as a metabolic enzyme inhibitor and as an adjuvant through its effect on enhanced cuticular penetration of deltamethrin 57 . The fact that these new generation nets (Type 1 and Type 3) were able to inhibit blood feeding more than current nets (Type 2 and Type 4), could suggest their capability to confer high personal protection against resistant mosquito biting.
Studies conducted in Benin and other African countries showed a loss efficacy of pyrethroids treated net against An. gambiae 9,23,50, 58 . This research has demonstrated that the efficacy of a combined pyrethroids-PBO nets on resistant malaria vector populations could be a promising strategy against pyrethroid resistance populations of Anopheles as previously highlighted 20,25,27,28,43,59,60 . This study further confirms a role of oxidases in pyrethroids resistance of An. funestus and the need to develop nets combining pyrethroid and synergist against pyrethroid resistant malaria vectors 25,33,40,61 . Nevertheless, several other studies have been conducted on the insecticide resistance management of Anopheles especially An. gambiae, using nonpyrethroid insecticides alone or in mixture of pyrethroids [62][63][64][65][66][67] . These studies also provided relatively good pointers for management of resistant mosquitoes but the problems with nonpyrethroid ingredients are human toxicity and their irritant effect 62,68 . A more recent study conducted by Malima et al. 69 in an area where An. funestus is resistant to pyrethroid, showed that even when non-pyrethroid insecticide-treated nets with durable wall lining (ITWL) are used, this cannot guarantee up to 50% protection against resistant An. funestus.

Conclusion
Pyrethroid resistance in the major malaria vectors An. funestus s.s. and An. coluzzii in Kpomè is high, and is likely to limit the impact of currently used LLINs. This study showed that the use of new generation bed nets could provide additional protection and reduce malaria burden in endemic environments. This study is of importance to Malaria control programs for improved control of pyrethroids resistant malaria vectors in Benin.

Ethical considerations
Approval was obtained from the ethics review boards of the International Institute of Tropical Agriculture (IITA) ref.PJ/ CC5339. All volunteers recruited to sleep in the experimental huts gave written and verbal consents. Chemoprophylaxis was provided to volunteer prior to the hut studies.

Data availability
All data generated and analyzed during this study will be included in the published article. Raw data are available from Open Science Framework. General remarks: The paper presents key results of an operational research topic with high significance for preventing malaria transmission in the context of multiple insecticide resistance among vector populations. The study is well designed and methods used are appropriate and linked with expected outcomes of the study.
Though, I noticed some limitations that have not been clearly sound for me (or confusing) while reading the whole manuscript.
For example: In Abstract: Comment 1: Key results presented do not meet with the objectives to achieve in the subsection "background": i) "bioefficacy of current LLINs" and ii) "performance of PBO LLINs in experimental hut". Key results on the second objective are missing.
Suggestion 1: provide key outcomes from experimental huts (at least mortalities by type of LLINs).
In Method section: Comment 2: the sequence of subsections as presented is confusing. The subsection "study design" in page 5 appears as heading for experiments in huts with a description of the five types of LLINs to evaluate. The description of such types of LLINS was made in relation with experimental huts and seems like a redundancy of the subsection "characteristics of LLINs ….." in page 4. But no confirmation if these 5 types of LLINS were different on not to those tested with cones or tunnels. The way you have calculate some key variables in tunnels (blood feeding inhibition, overall mortality), and in huts (personal protection, overall killing effect..) is not clear.
Suggestion 2: Confirm if types of LLINs (1 to 5) described in sections "chemical analysis of …" are also those used for all the bioassays including phase I (cone and tunnel tests) and phase II (experimental huts).
-If so (the 5 types of LLINs related as arms for all bioassay experiments), the study design may be moved after the "study area" and adjusted accordingly prior the description of "mosquito collection", "species identification", "susceptibility tests with insecticide paper and with PBO moved after the "study area" and adjusted accordingly prior the description of "mosquito collection", "species identification", "susceptibility tests with insecticide paper and with PBO synergist", "cone tests", "tunnel tests", "experimental huts" and "data analysis". Under "experimental hut" subsection may consider subparts like blank assessment of hut effects and release/recapture experiments.
-If no (the 5 types of LLINs differ between tests), clearly indicate as "experimental hut design" and keep at this position.
-Provide formula (available in WHOPES guidelines 2013) for the estimation of key parameters of LLIN bioefficacy in tunnels (blood feeding inhibition, overall mortality) and experimental huts (personal protection, overall killing effect..).
In Result section: Comment 3: All key results related to methods used have been presented. However, redundant data and information should be limited or avoided to shorter the paper, for example: Suggestion 3: - Table 1 (distribution of members of ….) in page 6 is not necessary (optional), since data have been considered in the text at the subsection "molecular speciation …"; - Figure 1 in page 6 does not add value since same data are considered in figure 2 in page 7. Consider only figure 2 and insert the specific data on Kisumu strain from figure 1 in the main text.
- Table 4 and 5 can be combined in one table of blank assessment of attractiveness and lethal effect of experimental huts (4 columns of variables per experimental hut: two for attractiveness: "overall mosquitoes" and "Anopheles" and two for lethal effect: "tested mosquitoes" and "mortality rate") -Check the last sentence of the third paragraph in page 8 (column on the right)…" Against resistant An. funestus s.s., blood feeding inhibition rates (removed "s") with net type 4 (41.75%) was significantly lower than net Type 4 (against? Should be type 3) (100%). Respectively...
-Titles of figures 2 (page 7) and 3 (page 8) may be more comprehensive: Figure 2. "Mortality rates of Anopheles to pyrethroids and when combined …" Figure 3: "Mortality rates reported with cone tests on different types of nets using pyrethroids resistant …" -Provide statistical evidences of the conclusion at the end of subsection "blank assessment of hut attractiveness" (page 9, 2nd paragraph in column on the left)..

In Conclusion section:
Comment 4: The conclusions of this paper are adequately supported by the results. Probably add elsewhere the justification and pertinence for undergoing phase II in experimental huts when LIN efficacy does not meet the WHO criteria using standard methods i.e. cone test (≥ 80% mortality or ≥ 95% knock-down) or tunnel tests (≥ 80% mortality or ≥ 90% blood-feeding inhibition).
Is the work clearly and accurately presented and does it cite the current literature?

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? Yes
Are all the source data underlying the results available to ensure full reproducibility? Partly

Are the conclusions drawn adequately supported by the results? Yes
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard. This manuscript describe results of experimental experiences performed to assess the efficacy of of pyrethroids/piperonyl butoxide (PBO) net treatments for controlling multi-resistant populations of . The study is highly interesting as deals with development of strategies of malaria Anopheles funestus s.s control in the context of high resistance to pyrethroid observed in main vectors. The manuscript is well written and presented, however there are very few minor concerns I would like the authors to take into consideration so that their work will be more efficient. :

Comments
The title could be modified to also take into account as all the results of the manuscript An. coluzzi are also presenting data from this species 2 paragraph of the introduction, in the 5 sentence : "resistance to insecticide ofone of ………" replace "ofone" by "of on" In the section "Methods", on the part "WHO cone tests with the nets" authors must keep the same nd th