Experiments were conducted with Anopheles gambiae sensu lato (s.l.) mosquitoes collected around Bouaké, central Côte d’Ivoire. This mosquito population has a high prevalence of resistance to the major classes of mosquito adulticides, including pyrethroids 36–38. Mosquitoes were collected as larvae from breeding sites using the dipping method and reared to adult in insectary under controlled temperature and humidity conditions (27 ± 2°C, 60 ± 20% RH). Larvae were fed on ground Tetramin baby fish food. Adult mosquitoes emerging from pupae were placed in 30cm x 30cm netted cages and maintained on 10% honey solution until testing.
Residual activity of beta-cyfluthrin treated EaveTube inserts under field conditions
This assessment was done as part of a cluster randomised controlled trial (CRT) in central Côte d’Ivoire. Forty villages were selected for the CRT with half assigned to household screening plus EaveTubes (SET) and the other half as controls 39. All villages received new LLINs, so the aim of the CRT was to investigate whether SET provides added protective benefit against malaria transmission on top of LLINs. Beta-cyfluthrin was selected for the CRT because this product was registered for use in country and results from a previous study indicated a long-lasting activity (>9 months) of this pyrethroid on electrostatic EaveTube inserts under controlled, semi-field conditions 16.
Inserts fitted to houses in the 20 intervention villages were machine-treated by In2care with a wettable powder formulation of 10% beta-cyfluthrin (Tempo 10©, Bayer). The dose of insecticide applied was in the range 300–500 mg per insert.
To monitor the efficacy of treated insert under field conditions in real houses, residual activity was tested monthly using a subsample of inserts from study villages using the eave tube bioassay.
The procedure of this bioassay was described in detail in Oumbouke et al. 16. In brief, the assay comprises of a 20-cm long plastic tube containing an insert such that it is flush with one end of the tube and mosquitoes are introduced into the tube through the opposite end, which is fitted with an untreated netting to keep mosquitoes inside the tube. A 1.5L plastic bottle filled with hot water and wrapped in socks worn the previous night was placed behind the insert and served as a host cue. Mosquitoes attracted to the heat and odour cues then contact the insecticide-laden insert. The eave tube assay is similar to the previously described MCD bottle assay 40 in that both mimic the interaction between host-seeking mosquitoes and insecticide-treated surfaces. To increase host-seeking activity, mosquitoes were starved for 6h prior to testing. Approximately 100 mosquitoes in batches of 20-25 were exposed for 1 hour in the eave tube bioassay. Following exposure, mosquitoes were released in netted cages and provided with 10% honey solution and mortality scored after 24h.
Four beta-cyfluthrin treated inserts were sampled from each EaveTube village every month for this monitoring activity. The number of inserts tested was based on logistical constraints in the field. Bioefficacy testing was performed monthly until activity decreased below 80% mortality at which point all of the inserts in the villages were replaced with freshly treated inserts.
Semi-field evaluation of two alternative insecticide delivery approaches in EaveTubes
Insecticide-treated electrostatic netting in tubes was shown to produce a significant reduction in overnight mosquito survival in previous semi-field studies 12,16−18. The experiments described here explore alternatives to electrostatic netting for delivering insecticides in this system. The following delivery methods were tested in experimental huts surrounded by enclosure (Fig. 1) at the M’bé field station near Bouaké, central Côte d’Ivoire:
In2Care EaveTube inserts (In2Care, the Netherlands) coated with a wettable powder formulation of 10% beta-cyfluthrin (Tempo 10©, Bayer) serves as a positive control. The dose of insecticide applied was in the range 300–500 mg per insert.
PermaNet 3.0 is a long-lasting insecticidal net manufactured by Vestergaard S.A. (Switzerland). The top panel, which was tested in the present study, is made of monofilament polyethylene (100 denier) fabric and treated with a mixture of the pyrethroid deltamethrin at 4g/kg and the synergist piperonyl butoxide (PBO) at 25g/kg. The side panels (not tested here) are made of multi-filament polyester (75 denier) fabric with a strengthened lower part incorporated with deltamethrin at 2.8g/kg.
Olyset Plus is a long-lasting insecticidal net manufactured by Sumitomo Chemical (Japan). The net is made of 150 denier high-density mono-filament polyethylene yarn incorporating a mixture of the pyrethroid permethrin at 20g/kg and PBO at 10g/kg on all net panels.
Interceptor G2 is a long-lasting net manufactured by BASF (Germany). The net is a dual-active LLIN made up of knitted multi-filament polyester fibres incorporating a mixture of the pyrethroid alpha-cypermethrin at 2.4g/kg and the pyrrole insecticide chlorfenapyr at 4.8g/kg.
The organophosphate pirimiphos methyl is a WHO recommended insecticide used extensively in IRS campaigns. Capsule suspension formulation of pirimiphos methyl (Actellic CS, Basel, Switzerland) was tested in the present study.
Semi-field performance of the alternative tube treatments was tested in two experimental huts at the M’bé field station, near Bouaké, central Côte d’Ivoire. The huts are of the West African design 41, 3.25 m long, 1.76 m wide and 2 m high. The interior walls of the huts are made of concrete brick, with a corrugated iron roof. A plastic cover was affixed onto the roofing as ceiling. Each hut was built on a concrete base with a water-filled moat, to prevent invertebrate predators from preying on dead or knocked down mosquitoes. A number of modifications were made to the huts for these experiments: (i) six holes were drilled at eave level (1.7 m from the ground) on three sides of the hut (two holes on each side), (ii) insecticide treated tubes were fitted into the holes, (iii) an enclosure was built around each hut to allow recapture of mosquitoes outside of the hut (Fig. 1). The semi-field enclosure consists of a wooden frame erected on the concrete base, 50 cm from the exterior wall of the hut. The roof was made of plastic sheeting which extended beyond the edge of the enclosure as an overhang to prevent rain from entering. The bottom half of the frame was made out of wooden panels and the top half was screened with polyethylene netting. White plastic sheeting was installed on the floor of the enclosure to facilitate the collection of dead mosquitoes. A zipper access door was positioned on the front side of the hut to allow entry into and exit from the enclosure.
In the first experiment, six 30 cm x 30 cm netting samples were cut from the LLINs and fitted in tubes in one experimental hut (the intervention). Six pieces of untreated netting of the same size were placed in the second experimental hut, located 50m away (the control). The netting samples were cut from Olyset Plus and Interceptor G2 and from the roof panel of PermaNet 3.0 and evaluated on different occasions.
In a second experiment, tubes were dipped in aqueous solution of pirimiphos methyl at 10g/m2. The tubes were treated by rolling one tube at a time in insecticide solution for 5 minutes and subsequently left to dry for 24h before testing. Tubes treated with pirimiphos methyl were screened with untreated netting (the intervention). A control hut fitted with untreated tube containing untreated netting was tested in parallel (the control). In the third experiment, six inserts freshly treated with beta-cyfluthrin were installed in one experimental hut (the intervention) and six untreated inserts were placed in tubes in a second experimental house (the control).
Two adult volunteers were recruited to sleep in the huts. Volunteer sleepers rotated between huts on consecutive nights to account for any potential difference in attractiveness to mosquitoes. The volunteers entered the hut at 20:00h and slept under intact untreated nets. Approximately 100 non-bloodfed sugar starved 5-day old female An. gambiae mosquitoes were released into each enclosure every release night 15 min after sleepers entered their respective huts. Mosquitoes were recaptured the following day at 05:00 inside the enclosure. Mosquitoes collected were brought back to the laboratory at Institut Pierre Richet (IPR) in Bouake, Côte d’Ivoire. Dead mosquitoes were counted and discarded. Surviving mosquitoes were provided with 10% honey solution and any delayed mortality was scored up to 72h later.
Sample size calculations
Evidence from previous semi-field studies suggests that insecticidal tube produces about 50% reduction in overnight mosquito survival 12,13,16,17. Based on this, the number of release night required to detect a 50% reduction in survival with 80% power and significance level of 5% was determined for each treatment in the R software using the “pwr” package. Eight replicates of release-recapture were performed for each treatment, which according to the sample size calculation was above the number required to demonstrate the expected effect size.
Insecticide susceptibility assays
Insecticide susceptibility assays were performed to measure susceptibility to the constituent actives in the LLINs and pirimiphos methyl in the local An. gambiae mosquito population. Diagnostic concentration of the pyrethroids deltamethrin (0.05%), permethrin (0.75%), alpha-cypermethrin (0.05%) and pirimiphos methyl (0.25%) were tested in WHO cylinders following WHO guidelines. A higher concentration of pirimiphos methyl (1%) was also tested in assays. Synergist assays were conducted by pre-exposing mosquitoes to PBO, which neutralises the activity of the cytochrome P450s involved in pyrethroid metabolism in mosquitoes. Because of stability issues with chlorfenapyr on filter paper, adapted Centre of Disease and Control (CDC) bottle bioassays were used to measure resistance to chlorfenapyr. Bottles were coated with chlorfenapyr at the diagnostic dose of 50µg/mL 42. Four replicates of 25 female mosquitoes (sugar fed, aged 2-3 days) were exposed for 1h to insecticide treated papers or bottles. Mortality was recorded 24h (pyrethroids) and 72h (chlorfenapyr) post-exposure. Mosquitoes in the control batch were held for 72h before scoring mortality.
Residual activity of new generation LLINs and pirimiphos methyl treatment
The residual activity of the best performing alternative delivery methods (PermaNet 3.0 roof and pirimiphos methyl coated PVC tube) in the release-recapture experiments was assessed.
Four 30cm x 30cm pieces from PermaNet 3.0 netting and four PVC tubes treated with pirimiphos methyl at the dosages of 1g/m2 and 10g/m2 were tested using the previously described eave tube assays 16. Testing was performed on the netting pieces and the treated tubes at monthly intervals. To evaluate AI decay under realistic conditions, the pieces of the LLINs (installed in tubes) and the IRS treated tubes were stored between testing in holes drilled at eave level in an experimental hut at the institute. Four replicates of 25 non-blood fed 6h sugar-starved, 5-day old mosquitoes were tested for each bioassay. Intervention and control mosquitoes were monitored for up to 72h before scoring post-exposure mortality.
When mortality decreased below 50%, the netting samples were washed once and re-tested in the eave tube bioassays. Net washing was conducted following WHO guidelines 43. Briefly, the pieces were washed individually for 10 min in a soap solution (savon de Marseille at 2g/L of deionised water) using a shaker bath set a 155 movements/min and 30°C. Samples were then rinsed twice in clean water for 10min and left to dry for 3-4 hours. Washed netting samples were tested only after full regeneration of the active ingredient (1 day) 44.
Content of deltamethrin and piperonyl butoxide was determined in the roof panel of unwashed PermaNet 3.0 netting at month 0, and the washed samples at month 2. Extraction of deltamethrin and PBO was performed using the CIPAC method 45. Both compounds were extracted by refluxing with xylene for 30 minutes in presence of dioctyl phthalate as internal standard and citric acid. Concentrations deltamethrin and PBO was subsequently measured by Gas Chromatography with Flame Ionization Detection (GC-FID).
All statistical analysis was performed using the R software version 3.5.3. Residual efficacy data across treatments was analysed using generalized linear models (GLMs) with the “arm” package. The models included insecticide treatments as independent variable and mosquito mortality as the outcome. Interactions between insecticides and residual efficacy testing interval were also included in the models. Pairwise comparisons were performed with the final model using the “multcomp” package. For the release-recapture experiments, generalized linear mixed models (GLMMs) with a binomial distribution and a logit link function was fitted to the data using the “lme4” package. The models included treatment as fixed effect. Enclosure, sleepers, and release-recapture study nights were treated as random effects. Significance of the fixed effect in the model was tested using likelihood ratio test (LRT). Susceptibility bioassay data were analysed using a χ 2-square test with Yates continuity correction.
Ethical clearance for the study was obtained from the ethics review committee of the London School of Hygiene and Tropical Medicine and the Côte d’Ivoire National Ethics Committee. Hut sleepers were all male and > 18 years old. Written informed consent was obtained from all volunteer sleepers taking part in the study prior to the release-recapture experiments.