Significant reduction in abundance of peridomestic mosquitoes (Culicidae) and Culicoides midges (Ceratopogonidae) after chemical intervention in western São Paulo, Brazil.

Background We assessed the impact of two sand fly insecticide interventions (insecticide spraying and insecticide-impregnated dog collars) on the peridomestic abundance and distribution of mosquitoes (Culicidae) and biting midges (Ceratopogonidae) in western Sao Paulo (Brazil) in a long-term (42-month) evaluation. Both of these dipteran groups are vectors of diseases of medical and veterinary relevance to humans and domestic animals in Brazil. Methods The interventions in the 3-arm stratified randomised control trial were: pheromone + insecticide (PI) (chicken roosts were sprayed with microencapsulated lambda-cyhalothrin; pheromone lure has no effect on the Diptera pests studied here); dog-collars (DC) (dogs fitted with deltamethrin-impregnated collars); and control (C) (unexposed to pyrethroids) were extended by 12 months. During that time, adult mosquitoes and midges were sampled along 280 households at three household locations (inside human dwellings, dog sleeping sites and chicken roosts). Results We collected 3145 culicids (9 genera, 87.6% Culex spp.) distributed relatively uniformly across all 3 arms: 41.9% at chicken roosts; 37.7% inside houses; and 20.3% at dog sleeping sites. We collected 11,464 Culicoides (15 species) found mostly at chicken roosting sites (84.7%) compared with dog sleeping sites (12.9%) or houses (2.4%). Mosquitoes and Culicoides were most abundant during the hot and rainy season. Increased daytime temperature was marginally associated with increased mosquito abundance (Z = 1.97, P = 0.049) and Culicoides abundance (Z = 1.71, P = 0.087). There was no significant association with daily average rainfall for either group. Household-level mosquito and midge numbers were both significantly reduced by the PI intervention 56% [incidence rate ratio, IRR = 0.54 (95% CI: 0.30-0.97), P ≤ 0.05] and 53% [IRR = 0.47 (95% CI: 0.26-0.85), P ≤ 0.05], respectively, compared to the control intervention. The abundance of both dipteran groups at dog sleeping sites was largely unaffected by the PI and DC interventions. The PI intervention significantly reduced abundance of mosquitoes inside houses (41%) and at chicken roosting sites (48%) and reduced midge abundance by 51% in chicken roosting sites. Conclusions Sprayed insecticide at chicken roosting sites reduced the abundance of mosquitoes and midges at the peridomestic level while dog collars had no effect on numbers for any group.