Efficacy of PermaNet® Dual compared to Interceptor® G2 and PermaNet 3.0 in experimental huts in Siaya County, western Kenya Authors: Nashon Ogutu, Silas Agumba, Vincent Moshi, Patrick Onyango, Collins Ouma, Edith Ramaita, Lenson Kariuki, John E. Gimnig, Bernard Abong’o, Eric Ochomo Published: 2024 ABSTRACT BackgroundPyrethroid-chlorfenapyr nets have shown significant epidemiological impact over pyrethroid-only and pyrethroid plus piperonyl-butoxide (PBO) in Africa. A non-inferiority evaluation of PermaNet® Dual, a new chlorfenapyr plus deltamethrin net, compared to Interceptor® G2, was conducted in experimental huts in Siaya, Kenya against free-flying pyrethroid-resistant Anopheles funestus.MethodsThis study was an experimental hut trial, following a 7 by 7 Latin Square design. Seven treatments and seven sleepers were deployed in the experimental huts daily and rotated weekly and daily, respectively. Mosquitoes were collected every morning between 06:30 h and 08:30 h and were assessed for blood feeding and then monitored for immediate knockdown 1-h post collection and delayed mortality after 72 h. Differences in proportional outcomes were analysed using the blocked logistic regression model, while differences in numerical outcomes were analysed using the negative binomial regression model. Non-inferiority determination was performed based on World Health Organization (WHO) protocol.ResultsMortality at 72 h was 30.2% for PermaNet 3.0, 44.4% for the Interceptor® G2 and 49.2% for the PermaNet® Dual. Blood feeding was highest with PermaNet® Dual at 15%, and least with PermaNet® 3.0 at 10%. PermaNet® Dual and Interceptor® G2 had no significant differences in mortality (OR = 1.10, 95% CI 1.00–1.20) or blood feeding (OR = 1.18, 95% CI 1.04–1.33) and the lower confidence bounds were within the non-inferiority margins but for blood feeding, non-inferiority was relatively high to the upper 95% confidence bound. PermaNet® Dual was non-inferior to the Interceptor® G2 and superior to the PermaNet® 3.0 nets in causing mortality but inferior to PermaNet ®3.0 in blood feeding inhibition of the vectors.ConclusionPermaNet® Dual met the WHO criteria for non-inferiority to Interceptor® G2 and may be considered for deployment for public health use against pyrethroid-resistant Anopheles vectors of malaria. Background Long-lasting insecticidal nets (LLINs) have contributed significantly to the decline in malaria transmission over the past two decades and remain the most widely used malaria vector control tool [1, 2]. LLINs provide a physical barrier against mosquito bites in addition to a toxic dose of insecticides which irritate, repel, knockdown and kill the mosquito resulting in reductions in blood feeding and reducing mosquito’s longevity overall [3, 4]. These chemical properties are beneficial as the LLIN ages and becomes holed [5]. Insecticide resistance threatens the effectiveness of these vector control tools [6, 7] and for this reason, there is a need for continuous innovation to ensure LLIN products remain effective against pyrethroid-resistant mosquitoes. Some of the approaches to mitigate insecticide resistance include the use of LLINs treated with a pyrethroid plus a synergist which is not directly toxic to mosquitoes but inhibits detoxification enzymes and restores susceptibility to insecticides. In September 2017, the World Health Organization (WHO) Global Malaria Programme released updated policy recommendations on the deployment of pyrethroid-PBO (piperonyl butoxide) LLINs [6] followed by the recommendation for deployment of PBO LLINs in areas of ongoing malaria transmission where the principal malaria vector(s) have developed pyrethroid resistance [7]. This recommendation was based on epidemiological data from cluster randomized control trials conducted with Olyset™ Plus in Tanzania which demonstrated that PBO LLINs have additional public health value [8]. Additional evaluations of pyrethroid PBO LLINs contributed further evidence of efficacy in the following years [9–12]. Since the recommendation of the first-in-class dual-active (dual-A.I.) LLINs after the demonstration of public health value in a community-based cluster randomized controlled trial (cRCT) [8, 13], many dual-A.I. LLINs such as Royal Guard® and Interceptor® G2 have been evaluated in WHO Phase I and II trials [14] and have shown promise compared to standard LLINs against pyrethroid-resistant vectors. Pyrethroid-PBO LLINs have faced three main challenges: First, PBO is a synergist for P450 monooxygenases, but insecticide resistance is often a result of additional mechanisms including metabolic, target-site [15], cuticular [16] and microbial factors [17]. Second, the PBO incorporated in some of these LLINs was observed to wane in concentration by 18–24 months, well before the expected lifetime of an LLIN which is assumed to be 36 months [18]. Third, the deployment of pyrethroid-PBO LLINs alongside IRS with organophosphates is potentially counterproductive as the P450 monooxygenases also serve to activate organophosphates into their toxic metabolites [19]. More recently, studies have demonstrated the additional benefits of dual active LLINs which incorporate pyrethroid and non-pyrethroid insecticides in the same net. The WHO has recently recommended two new classes of LLINs which combine pyrethroids and pyrroles, such as chlorfenapyr, and pyrethroids and insect growth regulators, such as pyriproxyfen (PPF) (https://www.who.int) [20] based on epidemiological impact. Chlorfenapyr acts by disrupting cellular respiration and oxidative respiration phosphorylation in mitochondria [21]. Its unique mode of action has potential for the control of pyrethroid-resistant mosquitoes [22, 23]. The Interceptor® G2 is a pyrethroid-chlorfenapyr LLIN developed by BASF, which has demonstrated improved control of pyrethroid-resistant malaria vectors in experimental hut trials in Benin, Burkina Faso and Tanzania [24–26]. Large-scale trials have also provided further evidence of epidemiological impact [27, 28]. Data from experimental hut studies are useful in comparing new products to first-in-class products that have epidemiological data supporting their use in non-inferiority trials [29]. This study evaluated the non-inferiority of PermaNet® Dual, a new pyrethroid-chlorfenapyr LLIN containing deltamethrin and chlorfenapyr against Interceptor® G2, a pyrethroid-chlorfenapyr LLINs containing alphacypermethrin and chlorfenapyr as well as the superiority of PermaNet® Dual over the PermaNet® 3.0 which contains deltamethrin and PBO. Methods Study site and experimental huts Experimental hut trials (EHTs) were conducted at the Dala Suna experimental hut site on the shores of Lake Kanyaboli (0° 02′ 08.5″ N, 34° 11′ 05.0″ E) in Alego Usonga sub-County, Siaya County, western Kenya. The huts are located close to the swamps that provide conducive breeding habitats for malaria vectors and are characterized by a high year-round abundance of Anopheles funestus and seasonal peaks of Anopheles arabiensis, with average household densities > 300 and > 20 per night, respectively [30]. The area experiences two rainy seasons, one from March to May and the other from October to November, with high malaria transmission throughout the year [31]. The primary economic activities of the local population are subsistence farming, livestock keeping, fishing and small-scale trading [30]. The experimental huts are designed to resemble a typical Kenyan household in structure and mosquito exit/entry points (eaves, windows and doors) (Fig. 1A). Mosquito exit traps were fitted to all four windows of the experimental huts, two windows on the front face and two on the backside of the huts. The walls of the huts are made of blocks and lined with mud on the inside. The floors are tiled with white tiles for easy collection of knocked-down and dead mosquitoes (Fig. 1B). The huts have corrugated iron roofs and a 10-cm eave gap. To prevent mosquitoes from exiting the huts, wood baffles are installed at the eave gaps, allowing easy entry for mosquitoes (Fig. 1C). Additionally, the huts are elevated above the ground on a concrete base surrounded by a water-filled moat to keep ants away [29].Fig. 1Experimental hut design: A front view of the hut fitted with window exit traps, B showing the tiled floor and the hut interior walls and C showing the wood baffles Experimental hut design: A front view of the hut fitted with window exit traps, B showing the tiled floor and the hut interior walls and C showing the wood baffles Baseline evaluation of insecticide resistance profile Larvae of Anopheles gambiae sensu lato (s.l.) were collected using the standard dipper (Model 320) from their natural breeding sites around the experimental hut site. Adult blood-fed An. funestus were collected indoors using the Prokopack (model 1419) from the houses surrounding the experimental hut site, after obtaining consent from household heads. The mosquito collection took place between August and October 2022. The collected mosquitoes were transported to the KEMRI-CGHR insectary, where An. gambiae s.l. larvae were raised into adults and blood-fed An. funestus were allowed to lay eggs, and their first filial (F1) generation reared to 3–5 days old adults for insecticide resistance testing. Larvae were reared in rainwater and fed on fine powder of Koi premium fish food under standard controlled conditions (27 ± 2 ºC, 80 ± 10% RH and 12:12 light-darkness). Upon emergence, adults were maintained on a 10% sugar solution until bioassay. The Kisumu strain of An. gambiae, an insecticide-susceptible strain, was also reared simultaneously under the same conditions and used as a bioassay control. To assess the susceptibility of An. gambiae s.l. and An. funestus from the Lake Kanyaboli experimental hut site to the active ingredients of insecticides in the LLINs to be tested, namely PermaNet 3.0 (deltamethrin + PBO) and Interceptor® G2(alpha-cypermethrin + chlorfenapyr) and PermaNet® Dual (deltamethrin + chlorfenapyr) LLINs, WHO tube assay and Center for Disease Control (CDC) bottle tests were conducted. WHO tube tests were carried out on 3 to 5-day-old F0 An. gambiae s.l. and F1 An. funestus adults according to the WHO protocol [32]. In brief, mosquitoes were exposed to filter papers impregnated with 0.05% alpha-cypermethrin, 0.75% permethrin or 0.05% deltamethrin for 1 h, during which knockdown was recorded every 10 min and mortality was recorded 24 h post-exposure. The intensity of insecticide resistance to pyrethroids was determined by increasing the diagnostic concentrations to 5X and 10X. non-insecticide-impregnated filter paper was also used as the control. The insecticide-treated filter papers were obtained from the WHO via the Kenya Medical Research Institute (KEMRI), and their quality was assessed against susceptible An. gambiae sensu stricto (s.s.) mosquitoes (Kisumu strain), as control. Pyperonyl butoxide (PBO) synergist test was also performed: mosquitoes were pre-exposed to 4% PBO for one hour and then exposed to 0.05% deltamethrin, 0.05% alpha-cypermethrin and 0.75% permethrin [32]. CDC Bottle bioassays were performed using the discriminating concentration of chlorfenapyr (100 µg/bottle) and clothianidin (4 µg/bottle + Mero) using both An. gambiae s.l and An. funestus following the WHO protocol [33]. Each Wheaton 250 ml bottle and its cap was coated with 1 ml of insecticide solution by rolling and inverting the bottles. In parallel, a control bottle was coated with 1 ml of acetone, followed by all bottles being covered with a sheet and left to dry overnight in the dark. Mosquitoes were exposed to chlorfenapyr and clothianidin for 60 min. Following exposure, mosquitoes were transferred to a netted paper cup, provided with lightly moistened cotton wool containing 10% sugar solution (changed daily) for chlorfenapyr-exposed mosquitoes and monitored at 24 h, 48 h, and 72 h. In contrast, clothianidin-exposed mosquitoes were monitored for 24 h only. Net treatments and treatment arms Both PermaNet® 3.0 and PermaNet® Dual were supplied by Vestergaard Sarl (Lausanne, Switzerland). Interceptor® G2 was supplied by BASF (Ludwigshafen, Germany). PermaNet® 3.0 was used in this evaluation as a comparator because it is the first dual-active insecticide-treated bed net Vestergaard S.a.r.l incorporating PBO on the top panel and deltamethrin on the side panels. It is also currently the standard of care being deployed in the area to combat pyrethroid resistance. The untreated nets were made of polyester fabric without any insecticide treatment. The Interceptor® G2 was made of polyester fabric coated with 2.4 g/kg (100 mg/m2) of alpha-cypermethrin and 4.8 g/kg (200 mg/m2) of chlorfenapyr. PermaNet® 3.0 was made of polyester fabric coated with 2.1 g/kg (84 mg/m2) of deltamethrin on the sides, and polyethylene incorporated with 4.0 g/kg (120 mg/m2) of deltamethrin and 25.0 g/kg (800 mg/m2) of PBO on the roof. PermaNet® Dual was made of polyester fabric coated with chlorfenapyr at 5.0 g/kg (200 mg/m2), and deltamethrin at 2.1 kg (84 mg/m2). Net washing For each study arm, Seven nets were randomly selected from a cohort of 21 nets of each production batch and subjected to twenty washes following the WHO washing criteria [14]. To prevent contamination between different nets, each LLIN type was washed separately in its washing station, which was equipped with separate assortments. The washing process involved immersing each net individually in a 16-L aluminium basin filled with 10 L of clean groundwater (pH of 7.0 and a hardness of 5 degrees), to which 20 g of soap was added and fully dissolved just before washing. Each net was washed for 10 min with agitation for 3 min, then soaked for 4 min and stirred again for 3 min. The net samples were rinsed twice in 10 L of clean groundwater using the same washing procedure, then dried under shade and stored at ambient temperature between washes. To simulate the wear-and-tear of the nets during use, all the LLINs intended for the hut trial of both treatment wash points and control nets were given 6 holes measuring 4 × 4 cm. Two holes were created on each of the long side panels and one hole on each of the short side panels, as per WHO guidelines [14]. Hut trial procedure Experimental hut trials used a 7 by 7 Latin square design (LSD) to evaluate the entomological efficacy of PermaNet® Daul, Interceptor® G2 and PermaNet® 3.0 LLINs washed 20 times and unwashed against free flying pyrethroid-resistant An. funestus. At each wash point, the efficacy of these LLINs was compared to an untreated net as a negative control. The trial used 49 nets, fourteen nets of each LLIN type (7 replicates of unwashed and seven replicates of washed), except for the untreated/control net, which had seven nets. Seven consenting human volunteer sleepers slept in the huts from 8:30 PM to 6:30 AM daily throughout the trial period, and to account for individual attractiveness to mosquitoes, they were rotated daily between the huts using a simple 7*7 LSD. The nets were erected inside the experimental huts by tying the edges of the roof panel to nails fixed at the upper corners of the hut wall using string. Treatments were rotated between experimental huts weekly according to a Latin square design to control the hut position effect. In contrast, volunteers were rotated daily to control differences in individual host attractiveness to mosquitoes. Mosquito collections were performed for 7 days in each collection round; on the 8th day, the huts were cleaned and aired to prevent contamination and carry-over effects before the next rotation cycle. The following treatment arms were evaluated in each experimental hut trial:Untreated net (control)—7 replicates of nets unwashed.PermaNet® Daul—7 replicates of nets washed 20 times.PermaNet® Daul—7 replicates of unwashed nets.Interceptor® G2—7 replicates of nets washed 20 times.Interceptor® G2—7 replicates of washed nets.PermaNet® 3.0—7 replicates of nets washed 20 times.PermaNet® 3.0—7 replicates of unwashed nets. Untreated net (control)—7 replicates of nets unwashed. Untreated net (control)—7 replicates of nets unwashed. PermaNet® Daul—7 replicates of nets washed 20 times. PermaNet® Daul—7 replicates of nets washed 20 times. PermaNet® Daul—7 replicates of unwashed nets. PermaNet® Daul—7 replicates of unwashed nets. Interceptor® G2—7 replicates of nets washed 20 times. Interceptor® G2—7 replicates of nets washed 20 times. Interceptor® G2—7 replicates of washed nets. Interceptor® G2—7 replicates of washed nets. PermaNet® 3.0—7 replicates of nets washed 20 times. PermaNet® 3.0—7 replicates of nets washed 20 times. PermaNet® 3.0—7 replicates of unwashed nets. PermaNet® 3.0—7 replicates of unwashed nets. Mosquito collections and processing Seven consenting human volunteers slept in experimental huts from 8:30 PM to 06:30 AM during each trial to attract wild, free-flying mosquitoes. All the sleepers were provided with weekly prophylaxis (mefloquine) and instructed to record any side effects experienced during the evaluation period. From 6:30 AM, mosquito collections were conducted using mouth aspirators until 08:00 each morning. The sleepers collected all the dead and alive mosquitoes inside the huts and window exit traps using mouth aspiration. The mosquitoes were scored based on their point of collection, such as wall, roof, floor, net, and under-bed, as well as from the window exit traps. Once collected, the mosquitoes were transferred into clean, netted paper cups and provided with access to a 10% sugar solution. The samples were arranged in cooler boxes and transported to the field insectary laboratory. In the laboratory, the mosquitoes were sorted by status (alive or dead: blood-fed or unfed: gravid or half-gravid) and identified morphologically to species following taxonomical key [34]. All the live mosquitoes were observed for knockdown one-hour post collection, and mortality was recorded every 24 h for 72-h. Supplementary laboratory assays Cone test Cone testing was performed with net pieces (25 cm × 25 cm) drawn from before and after the field trial of all wash points of all LLINs used in this evaluation. Four cones were attached to each net piece, and five non-blood-fed female mosquitoes were aspirated into each of the four cones and exposed for 3 min [14]. Both An. gambiae, Kisumu strain and An. funestus F1 of 3–5 days were introduced in each cone. In total, 100 mosquitoes were used per net/species. After exposure, the mosquitoes were transferred into clean paper cups, provided with a 10% sugar solution, and knockdown was recorded 60 min post-exposure, with mortality recorded at 24 h, 48 h, and 72 h. Mosquitoes were kept under the same laboratory conditions described above. The insecticide-susceptible strain of An. gambiae Kisumu was used as control. Tunnel test The tunnel test measures host-seeking mosquitoes' mortality and blood-feeding success in an experimental chamber. This experiment was designed to provide further insight and explain the toxicity of unwashed and washed nets used in the huts. Tunnel assays were conducted against the pyrethroid-resistant An. funestus F1 from the experimental hut site with the same net pieces of Interceptor® G2 and PermaNet® Dual tested in the cone assay. The tunnel test chamber mimics the behavioural interactions between free-flying mosquitoes and nets during host-seeking. It consists of a square glass tunnel divided one-third (20 cm) of its length by a box frame fitted with a net sample Fig. 2. In the short section of the tunnel, a rabbit bait was held in a cage with its back sheared and exposed for easy accessibility and feeding by mosquitoes [14]. In contrast, in the long sections (40 cm), 100 5–8-day-old mosquitoes were released at 6:00 PM and left until 7:00 AM under standard controlled conditions (27 ± 2 ºC temperature, 80 ± 10% RH). The net pieces used in the experiment had nine small holes, each measuring 1 cm in diameter, which allowed mosquitoes to enter the baited chamber. The mosquitoes were collected from the tunnel in the morning and examined for mortality and blood-feeding success. The surviving mosquitoes were placed in clean paper cups with a label and given access to a 10% sugar solution. Delayed mortality of the live mosquitoes was recorded every 24 h, up to a maximum of 72 h.Fig. 2A Tunnel assay set up in the laboratory to assess mosquito mortality and blood feeding success with dual active LLINs A Tunnel assay set up in the laboratory to assess mosquito mortality and blood feeding success with dual active LLINs Chemical assays Two nets were randomly selected from all the wash points in every arm, before and after the hut trials, and five pieces were obtained from each net apart from PermaNet 3.0, from which 3 pieces were obtained from the top and 1 from each side (7 pieces total) following WHO guidelines on net cutting. The cut net pieces were shipped wrapped in aluminium foil to the Vestergaard ISO/IEC 17025 accredited Vector Control Laboratories in Vietnam for testing to determine the wash retention of active ingredients in the net pieces using analytical methods validated and published by the Collaborative International Pesticides Analytical Council (CIPAC). Briefly, deltamethrin in the roof of PermaNet 3.0 (roof) was extracted from net samples by heating under reflux for 30 min with xylene using dicyclohexyl phthalate as internal standard. The solvent was evaporated, and the residue dissolved in hexane. Deltamethrin was extracted from the nets, including PermaNet® Dual and PermaNet 3.0 sides using dicyclohexyl phthalate and the concentration was determined by normal phase high-performance liquid chromatography with UV diode array detection (HPLC–DAD). Alpha-cypermethrin in Interceptor® G2 as well as chlorfenapyr in Interceptor® G2 and PermaNet® Dual were sonicated with heptane using dicyclohexyl phthalate as internal standard and determined by gas chromatography with flame ionisation detection (GC-FID). Lastly, PBO in PermaNet 3.0 roof was extracted from net samples by heating under reflux for 30 min with xylene using octadecane as internal standard and determined by GC-FID. Data analysis The primary outcomes measured by comparing the treatments and control experimental huts were blood-feeding inhibition (the reduction in blood feeding in treatments compared with that in the control huts), immediate and delayed mortality (the proportion of mosquitoes that are dead in the morning of collection and the cumulative proportion dead at 24, 48 or 72 h). In addition, induced exophily (the proportion of mosquitoes that are found in the exit traps) and deterrence (proportional reduction in the number of mosquitoes collected in the treated huts relative to the number collected in the control huts with untreated nets) were evaluated. The difference in proportional outcomes (mortality, blood feeding and exophily) between treatments and control at all wash points were analysed using a blocked logistic regression model, while differences in numerical outcomes (entry) were analysed using a negative binomial regression model. Tests of non-inferiority between PermaNet® Dual and Interceptor® G2 for both mortality and blood feeding were performed according to the WHO protocol [35]. The analysis included both washed and unwashed nets with an independent variable in the washing model. A candidate product is considered non-inferior to the active comparator product if: (a) the lower 95% confidence interval of the odds ratio describing the difference in mortality between the candidate and comparator product is > 0.7 and/or, (b) the upper 95% confidence interval of the odds ratio describing the difference in blood feeding between the candidate and comparator product is < 1.43. The superiority between PermaNet® Dual and PermaNet 3.0 was also assessed based on whether mortality rates were higher and blood feeding rates lower at a 5% significance level (i.e. p < 0.05). All analyses were done using R Statistical Software (v4.2.2; R Core Team 2021). Ethical considerations and compliance with GLP Ethical approval for the trial was issued by the Scientific and Ethical Review Unit of KEMRI (SERU 4536) for involving humans and animals. This study was also reviewed by the CDC and was determined to meet the definition of research involving human subjects. Still, the CDC’s involvement was not considered to constitute an engagement in human subjects research. Prior to recruitment into the study, formal informed consent was obtained from the volunteer sleepers. The participants were each given a weekly course on malaria prophylaxis (Mefloquine) to protect them from contracting malaria. This site is accredited by the Kenya Pest Control Products Board (PCPB) for the national evaluation of vector control products for registration purposes. The study was conducted in strict conformance with WHO non-inferiority guidelines for the evaluation of second-in-class LLINs [35]. Additionally, the site has begun the process towards GLP accreditation and conducts all study procedures in strict conformance with GLP requirements. Results Insecticide resistance profile of the local mosquito populations No mortality was recorded in the controls. Therefore, Abbott's formula was not used to correct the mortality rates. Pyrethroid resistance was detected in all species (Table 1). An. gambiae s.l. and An. funestus from the study area showed resistance to the diagnostic dose of deltamethrin (1X), with only 45% and 72% mortality observed, respectively. Although there was an increase in mortality when exposed to higher doses of deltamethrin (5X), the mortality rates increased from 45 and 72% (for the 1X dose) to 84% and 77% for An. gambiae s.l. and An. funestus, respectively. Exposure to the highest 10X diagnostic dose resulted in 100% mortality for An. gambiae and 92% for An. funestus. The results were comparable for both species when testing permethrin and alpha-cypermethrin insecticides, with none achieving 100% mortality even after increasing the diagnostic doses to 10 times the standard dose. Pre-exposure to PBO restored full susceptibility to deltamethrin and partial restoration of susceptibility to permethrin and alpha-cypermethrin in the An. gambiae population, but susceptibility was partially restored in An. funestus to all tested pyrethroids. The non-pyrethroids insecticides (pirimiphos methyl, clothianidin and chlorfenapyr) tested using CDC bottle bioassay resulted in 100% mortality when exposed to diagnostic doses.Table 1Insecticide resistance status of malaria vectors of Lake Kanyaboli, western KenyaAssaysInsecticideDoseConcentrationSample size% MortalityAn. gambiaeAn. funestusWHO tubeAlphacypermethrin1X0.05%10082455X0.25%100886010X0.50%1009394PBO + Alphacypermethrin1X0.05%1009597Deltamethrin1X0.05%10045775X0.25%100847210X0.50%10010092PBO + deltamethrin1X0.05%10010097Permethrin1X0.75%10082645X3.75%100989410X7.50%10010086PBO + permethrin1X0.75%1009995Pirimiphos-methyl1X0.25%100100100WHO bottleClothianidin4 µg/ml100100100Chlorfenapyr100 µg/ml100100100 Table 1Insecticide resistance status of malaria vectors of Lake Kanyaboli, western KenyaAssaysInsecticideDoseConcentrationSample size% MortalityAn. gambiaeAn. funestusWHO tubeAlphacypermethrin1X0.05%10082455X0.25%100886010X0.50%1009394PBO + Alphacypermethrin1X0.05%1009597Deltamethrin1X0.05%10045775X0.25%100847210X0.50%10010092PBO + deltamethrin1X0.05%10010097Permethrin1X0.75%10082645X3.75%100989410X7.50%10010086PBO + permethrin1X0.75%1009995Pirimiphos-methyl1X0.25%100100100WHO bottleClothianidin4 µg/ml100100100Chlorfenapyr100 µg/ml100100100 Insecticide resistance status of malaria vectors of Lake Kanyaboli, western Kenya Mosquito entry and exit rates in experimental huts A total of 15,114 pyrethroid-resistant female An. funestus were collected during the experimental hut evaluation. More mosquitoes were collected in huts with the unwashed PermaNet® 3.0 compared to the washed PermaNet® 3.0 and the washed and unwashed Interceptor® G2. Exit rates were significantly higher for the washed and unwashed PermaNet® 3.0 compared to all other treatments, while the exit rates for the unwashed Interceptor® G2 were significantly lower than the untreated net. No other significant differences in exit rates were observed. Non-inferiority assessment from the experimental hut According to the recent provisional WHO guidelines, for a candidate LLIN to be included in an established intervention class, it must demonstrate non-inferiority to the first-in-class product which has already demonstrated public health value (Interceptor® G2, for pyrethroid-chlorfenapyr ITN class) and superiority to pyrethroid only LLIN in experimental hut trial [35]. The non-inferiority margin is set at 0.7 for mortality and 1.43 for blood feeding. The odds ratio for the difference in mosquito mortality between PermaNet® Dual and Interceptor® G2 was 1.21 (95% confidence interval 1.093587–1.337), while the odds ratio for the difference in mosquito blood feeding was 1.18 (95% confidence interval 1.04–1.33) in mosquitoes. Following the WHO criteria described above, PermaNet® Dual is non-inferior to Interceptor® G2 based on the mortality (49% vs 44%, p < 0.047) induced in pyrethroid-resistant An. funestus in the experimental hut trial in Lake Kanyaboli, Kenya, while the PermaNet® Dual is both inferior and non-inferior to the Interceptor® G2 based on blood-feeding inhibition (85% vs 87%, p < 0.001) (Table 2). For the superiority assessment, PermaNet® Dual was superior to PermaNet® 3.0 in mortality induced (49% vs 30%, p < 0.001) but was inferior to PermaNet® 3.0 in blood feeding (10% vs 15%, p < 0.001) (Table 2). Due to the high control mortality (37%), Abbott’s correction was applied to all mortality data (Table 3). After correction, the mortality rates were 30.2%, for PermaNet® 3.0, 49.2% for PermaNet® Dual, and 44.4% for Interceptor® G2. Despite the reduction in absolute mortality rates, the relative performance of the nets and the conclusions regarding non-inferiority and superiority remained consistent with the uncorrected data.Table 2Results from the non-inferiority assessment of PermaNet® Dual to Interceptor® G2 against wild pyrethroid-resistant An. funestus in experimental huts in Siaya, western KenyaPrimary indicatorsVariablesPermaNet® 3.0PermaNet® DualInterceptor® G2PermaNet® DualTotal collected4390448140364481MortalityTotal dead2441302926293029Observed mortality (%)56686568Corrected mortality (%)30.249.244.449.2Odds ratio–2.246–1.211Std. error (on log odds scale)–0.080–0.050P-value– < 0.001–0.04795% CIs–1.920–2.627–1.097–1.337WHO efficacy criteria–Significantly higher (p < 0.05)–Lower 95% CI > 0.7Conclusion–Superior–Non-inferiorBlood feedingTotal blood-fed423671517671Blood-feeding (%)10151315Odds ratio–1.627–1.176Std. error (on log odds scale)–0.110–0.076P-value– < 0.001–0.01295% CIs–1.425–1.856–1.037–1.334WHO efficacy criteria–Significantly lower (p < 0.05)–Upper 95% CI < 1.43Conclusion–Inferior–Inferior and non-inferiorTable 3Results from the non-inferiority assessment of PermaNet® Dual to Interceptor® G2 and superiority to PermaNet® 3.0 against wild pyrethroid-resistant An. funestus in experimental huts in Siaya, western Kenya (with Abbott’s correction applied to mortality data and regression analysis results included)OutcomeTreatmentTotal collectedEvent count%Corrected %Regression coefficient (SE)Odds ratio (95% CI)P-valueMortalityControl220781737.0–Reference––PermaNet® 3.04390244155.630.20.809 (0.054)2.246 (2.021–2.495) < 0.001PermaNet® Dual4481302967.649.21.304 (0.053)3.684 (3.320–4.088) < 0.001Interceptor® G24036262965.144.41.167 (0.054)3.212 (2.889–3.572) < 0.001Blood feedingControl220792642.0–Reference––PermaNet® 3.043904239.6–− 1.869 (0.068)0.154 (0.135–0.176) < 0.001PermaNet® Dual448167115.0–− 1.407 (0.062)0.245 (0.217–0.276) < 0.001Interceptor® G2403651712.8–− 1.573 (0.065)0.207 (0.183–0.235) < 0.001Control mortality was 37%Abbott's correction was applied to mortality dataRegression coefficients and odds ratios are from blocked logistic regression models adjusting for hut, sleeper, day, and wash status Table 2Results from the non-inferiority assessment of PermaNet® Dual to Interceptor® G2 against wild pyrethroid-resistant An. funestus in experimental huts in Siaya, western KenyaPrimary indicatorsVariablesPermaNet® 3.0PermaNet® DualInterceptor® G2PermaNet® DualTotal collected4390448140364481MortalityTotal dead2441302926293029Observed mortality (%)56686568Corrected mortality (%)30.249.244.449.2Odds ratio–2.246–1.211Std. error (on log odds scale)–0.080–0.050P-value– < 0.001–0.04795% CIs–1.920–2.627–1.097–1.337WHO efficacy criteria–Significantly higher (p < 0.05)–Lower 95% CI > 0.7Conclusion–Superior–Non-inferiorBlood feedingTotal blood-fed423671517671Blood-feeding (%)10151315Odds ratio–1.627–1.176Std. error (on log odds scale)–0.110–0.076P-value– < 0.001–0.01295% CIs–1.425–1.856–1.037–1.334WHO efficacy criteria–Significantly lower (p < 0.05)–Upper 95% CI < 1.43Conclusion–Inferior–Inferior and non-inferior Results from the non-inferiority assessment of PermaNet® Dual to Interceptor® G2 against wild pyrethroid-resistant An. funestus in experimental huts in Siaya, western Kenya Table 3Results from the non-inferiority assessment of PermaNet® Dual to Interceptor® G2 and superiority to PermaNet® 3.0 against wild pyrethroid-resistant An. funestus in experimental huts in Siaya, western Kenya (with Abbott’s correction applied to mortality data and regression analysis results included)OutcomeTreatmentTotal collectedEvent count%Corrected %Regression coefficient (SE)Odds ratio (95% CI)P-valueMortalityControl220781737.0–Reference––PermaNet® 3.04390244155.630.20.809 (0.054)2.246 (2.021–2.495) < 0.001PermaNet® Dual4481302967.649.21.304 (0.053)3.684 (3.320–4.088) < 0.001Interceptor® G24036262965.144.41.167 (0.054)3.212 (2.889–3.572) < 0.001Blood feedingControl220792642.0–Reference––PermaNet® 3.043904239.6–− 1.869 (0.068)0.154 (0.135–0.176) < 0.001PermaNet® Dual448167115.0–− 1.407 (0.062)0.245 (0.217–0.276) < 0.001Interceptor® G2403651712.8–− 1.573 (0.065)0.207 (0.183–0.235) < 0.001Control mortality was 37%Abbott's correction was applied to mortality dataRegression coefficients and odds ratios are from blocked logistic regression models adjusting for hut, sleeper, day, and wash status Results from the non-inferiority assessment of PermaNet® Dual to Interceptor® G2 and superiority to PermaNet® 3.0 against wild pyrethroid-resistant An. funestus in experimental huts in Siaya, western Kenya (with Abbott’s correction applied to mortality data and regression analysis results included) Control mortality was 37% Abbott's correction was applied to mortality data Regression coefficients and odds ratios are from blocked logistic regression models adjusting for hut, sleeper, day, and wash status Supplementary assay results Both washed and unwashed PermaNet® Dual and Interceptor® G2 pieces tested induced low mortality in cone bioassays (< 73% for all tests, Fig. 3) against susceptible An. gambiae s.s, Kisumu strain, indicates that the cone bioassay is unsuitable for testing slow-acting actives even when combined with pyrethroids, a fast-acting active ingredient. PermaNet 3.0 roof net pieces induced the highest mortality rates (100%) for all the wash points, with sides-inducing mortality rates of > 92% (Fig. 4).Fig. 3Cone assay mortality results of An. gambiae, Kisumu strain when exposed to PermaNet® 3.0 net for 3 min. Error bars represent 95% confidence intervalsFig. 4Cone assays mortality result of An. gambiae, Kisumu strain when exposed to dual actives ITNs following WHO guidelines. Error bars represent 95% confidence intervals Cone assay mortality results of An. gambiae, Kisumu strain when exposed to PermaNet® 3.0 net for 3 min. Error bars represent 95% confidence intervals Cone assays mortality result of An. gambiae, Kisumu strain when exposed to dual actives ITNs following WHO guidelines. Error bars represent 95% confidence intervals Tunnel assays results Mortality rates of An. funestus in tunnel tests against the Interceptor® G2 and the PermaNet® Dual were high at all wash points (> 96.6%). Interceptor® G2 induced the highest mortality rate with 20 washes after the hut trial at 99.1% while PermaNet® Dual had the highest mortality rate of 98.2% with unwashed net pieces obtained from LLINs pieces after the hut trial. However, there was no significant difference in mortality between the two pyrethroid–chlorfenapyr LLINs (Fig. 5).Fig. 5Mortality rate of pyrethroid-resistant An. funestus F1 mosquitoes exposed to Interceptor® G2, PermaNet® Dual and PermaNet® 3.0 in tunnel tests. Error bars represent 95% confidence intervals Mortality rate of pyrethroid-resistant An. funestus F1 mosquitoes exposed to Interceptor® G2, PermaNet® Dual and PermaNet® 3.0 in tunnel tests. Error bars represent 95% confidence intervals High blood-feeding inhibition of 96% was witnessed with samples of unwashed PermaNet® Dual after the hut trial whereas Interceptor® G2washed 20 times pieces after the hut trial induced the lowest blood-feeding inhibition of 80% (Fig. 6).Fig. 6Blood-feeding inhibition of pyrethroid-resistant An. funestus, Siaya strain against new generation nets in tunnel tests. The red lines indicate WHO cut-off criteria for efficacy in tunnels Blood-feeding inhibition of pyrethroid-resistant An. funestus, Siaya strain against new generation nets in tunnel tests. The red lines indicate WHO cut-off criteria for efficacy in tunnels Chemical assays All the unwashed LLINs had AI content within the manufacturer-specified range. Retention of AI was lowest in the net pieces cut from the PermaNet® Dual washed 20 times (43% deltamethrin and 47% chlorfenapyr) and highest in the net pieces cut from the Interceptor® G2 washed 20 times (83.5% alpha-cypermethrin and 81% chlorfenapyr). Net pieces obtained from the PermaNet® 3.0 had retention of 64, 92.8 and 82% for deltamethrin on the sides, deltamethrin on the roof and PBO, respectively (Table 4).Table 4The content of active ingredients contained in unwashed and washed net pieces before and after the experimental hut trial in Siaya, KenyaITN BrandActive ingredient (s)AI content (g/kg)AI retention (%)UnwashedWashed 20XPermaNet® 3.0Deltamethrin (sides)1.751.1264.0Deltamethrin (roof)3.613.3592.8PBO (roof)19.1115.6982.1Interceptor® G2Alpha-cypermethrin2.852.3883.5Chlorfenapyr5.564.5181.1PermaNet® DualDeltamethrin2.090.9043.1Chlorfenapyr5.002.3847.6 Table 4The content of active ingredients contained in unwashed and washed net pieces before and after the experimental hut trial in Siaya, KenyaITN BrandActive ingredient (s)AI content (g/kg)AI retention (%)UnwashedWashed 20XPermaNet® 3.0Deltamethrin (sides)1.751.1264.0Deltamethrin (roof)3.613.3592.8PBO (roof)19.1115.6982.1Interceptor® G2Alpha-cypermethrin2.852.3883.5Chlorfenapyr5.564.5181.1PermaNet® DualDeltamethrin2.090.9043.1Chlorfenapyr5.002.3847.6 The content of active ingredients contained in unwashed and washed net pieces before and after the experimental hut trial in Siaya, Kenya Discussion This study evaluated the efficacy (mortality and blood feeding inhibition) and wash resistance of PermaNet® Dual (Vestergaard) in comparison to Interceptor® G2 (BASF) and PermaNet® 3.0. (Vestergaard) against pyrethroid-resistant free-flying An. funestus mosquitoes in experimental huts on the shores of Lake Kanyaboli in Siaya County, western Kenya. This locality has a year-round abundance of An. funestus and seasonal abundance of An. arabiensis. This trial was conducted in the dry season and therefore only An. funestus had adequate numbers for statistical comparisons, averaging 44 female mosquitoes per hut per night. The Lake Kanyaboli area is mostly swampy with permanent stagnated pools of water conducive to the development of An. funestus s.s. with peak numbers > 300 mosquitoes per structure per night in the rainy seasons [30]. All three LLINs evaluated here had significantly higher mortality rates on the free-flying An. funestus mosquitoes relative to the control in the experimental huts. PermaNet® Dual induced the highest mortality rates which was not significantly different from Interceptor® G2 but was significantly higher than PermaNet® 3.0 at corroborating results from previous hut trials in Benin [36]. Similar observations have been made in experimental hut trials evaluating PermaNet® Dual and Interceptor® G2 where in each instance, the pyrethroid-chlorfenapyr LLIN induced higher mortality than the pyrethroid-PBO or pyrethroid-only LLINs [25, 36, 37]. The application of Abbott's correction to account for high control mortality resulted in lower absolute mortality rates for all treatments. However, the relative efficacy of the nets remained consistent, with PermaNet® Dual still demonstrating non-inferiority to Interceptor® G2 and superiority to PermaNet® 3.0 in terms of mosquito mortality. This suggests that while environmental factors may have influenced overall mosquito survival in the experimental huts, they did not substantially alter the comparative performance of the different net types. Despite the lower mortality observed in this and other experimental hut studies, pyrethroid-PBO LLINs have been shown to offer up to 2 years better protection, with reduced parasite prevalence and vector densities than pyrethroid-only LLINs in Uganda and Tanzania [8, 9]. However, the rapid loss of PBO is a concern. A study in Tanzania noted that the PBO content of the nets was significantly reduced at 12 months and was almost lost by 24 months, a risk for sustained efficacy against pyrethroid-resistant malaria mosquitoes over the expected three-year lifetime [10]. For this reason, dual active nets with three years of effectiveness are urgently needed to complement vector control efforts in areas of high pyrethroid resistance. High resistance to alphacypermethrin, deltamethrin and permethrin was observed in both An. funestus and An. arabiensis which coincides with earlier reports [38, 39]. Higher concentrations of deltamethrin and permethrin in WHO tube assays (0.50% and 7.5%, respectively) and deltamethrin and alpha-cypermethrin (5X and 10X, respectively) in bottle assays were effective against An. gambiae, but not against An. funestus, indicating a higher intensity of resistance in An. funestus relative to sympatric vectors. Full susceptibility of both malaria vectors from the area to non-pyrethroids insecticides at standard doses: neonicotinoids (clothianidin), pyrrole (chlorfenapyr) and organophosphate (pirimiphos-methyl) was observed, despite high resistance to pyrethroids indicating that these classes could be effective for rotation or use of mixture formulations for malaria control in the region. The above finding was also an indication that there was no cross-resistance between pyrethroids and these other classes of insecticides. The addition of PBO as a synergist was observed to partially restore the observed susceptibility in both An. arabiensis and An. funestus indicating the involvement of P450 monooxygenases in the resistant phenotypes as has been reported elsewhere [40–42]. However, it partially restored susceptibility to > 95% mortality, which is close to full susceptibility, which suggests the involvement of other resistance mechanisms. PermaNet® Dual was non-inferior to Interceptor® G2 (the first in class), with an odds ratio of 1.21 (1.10–1.34, P > 0.05) at a non-inferiority margin of 0.7 according to the WHO guidelines for evaluation of non-inferiority to first in class products [43]. Following this criterion, PermaNet® Dual does not need to undergo evaluation for epidemiological impact but is available for recommendation as a second product in the same class. The PermaNet® Dual has since been prequalified by the WHO (https://extranet.who.int/pqweb/vector-control-product/PermaNet-dual) and is therefore available for immediate deployment to contribute to insecticide resistance management (IRM). Additionally, PermaNet® Dual was superior in inducing mortality relative to PermaNet® 3.0 with an odds ratio of 2.25 (1.92–2.63, P > 0.001). This shows the contribution of chlorfenapyr to the control of resistant mosquitoes where mechanisms other than P450 monooxygenases are active such as in this population. Similar findings have been documented in Tanzania [10], where there was a higher impact on entomological outcomes in clusters with Interceptor® G2 than those with PermaNet® 3.0, and in another experimental hut trial evaluating the non-inferiority of PermaNet® Dual to Interceptor® G2 [36]. Blood-feeding inhibition was significantly higher with PermaNet® 3.0 compared to both Interceptor® G2 and PermaNet® Dual but was not significantly different between the two pyrethroid-chlorfenapyr nets. Results from a separate study comparing ® G2 and chlorfenapyr-only control showed higher blood-feeding rates in the chlorfenapyr-only arm indicating that pyrethroids contribute the most to blood-feeding inhibition [37]. The current study indicates that PBO in PermaNet® 3.0 synergized the blood-feeding inhibition and, therefore, lower blood-feeding rates were achieved compared to the pyrethroid-chlorfenapyr nets. PermaNet® 3.0 was superior to PermaNet® Dual in blood-feeding inhibition. PermaNet® Dual was non-inferior to Interceptor® G2 in blood feeding inhibition possibly due to the higher irritability of alpha-cypermethrin. These results were not significantly different between unwashed nets and nets washed 20X, although the trends were towards higher mortality in the 20X washed nets. PermaNet® Dual and Interceptor® G2 did not have reductions in induced mortality or blood feeding inhibition after 20 washes, indicating good wash resistance, which is the current standard WHO proxy for an LLIN giving good performance for up to three years of use, despite less than 50% AI retention in the PermaNet® Dual. Previous studies have reported similar results [36, 37]. Standard laboratory cone bioassays with PermaNet® Dual and Interceptor® G2 failed to predict their efficacy against pyrethroid-resistant An. funestus s.l. in experimental huts. Cone bioassays with pyrethroid-chlorfenapyr nets did not meet the WHO criteria for the susceptible An. Gambiae, Kisumu strain PermaNet® 3.0 while tunnel tests with the PermaNet® Dual resulted in > 95% mortality against F1 progeny of wild An. funestus, affirming the unsuitability of cone bioassays for the evaluation of chlorfenapyr LLINs. These findings are similar to earlier ones reported in Benin and Cote d’Ivoire [36, 44] and indicate that tunnel tests are required as a laboratory assay of pyrethroid plus chlorfenapyr nets. The primary limitation of the study is the high mosquito mortality rates (37%) observed in the control huts. There was a significant difference in mosquito mortality rates between mosquitoes collected from the control exit trap (63%) and the control indoor (11%), averaging 37%. This suggests that the high mortality in the control hut could be attributed to strong and swift winds around the lake where the experimental hut is located. In addition, there was an unexpectedly high rate of exophily in the control arm which was higher than PermaNet® Dual and Interceptor® G2 at 28%, and which could not be explained. Mortality at 72 h in the control arm was 37% which was higher than most other hut studies including another pyrethroid-chlorfenapyr net experimental hut study with An. funestus in Tanzania [30]. This was likely due to excess mortality in the exit traps as the experimental hut sites are located on the shores of Lake Kanyaboli and receive strong winds through the night which desiccated the mosquitoes which escaped into the exit traps leading to increased mortality. However, given the high densities of An. funestus per hut per day (44), this did not affect the statistical power of the study. Conclusions PermaNet® Dual, the candidate product (deltamethrin + chlorfenapyr), was non-inferior to Interceptor® G2, the reference product (alphacypermethrin + chlorfenapyr) in causing mortality and inducing blood-feeding inhibition of free-flying wild pyrethroid-resistant An. funestus in this experiment. PermaNet® Dual was superior to PermaNet® 3.0, the positive control (deltamethrin + PBO) in causing mortality but inferior in the blood-feeding inhibition of wild pyrethroid-resistant An. funestus in this experiment. Overall, PermaNet® Dual met the WHO efficacy criteria in relation to non-inferiority to Interceptor® G2 and can be deployed in areas of high pyrethroid resistance.