Methods to collect Anopheles mosquitoes and evaluate malaria transmission: A comparative study in two villages in Senegal

Background: Various methods have been studied as replacement of human landing catches (HLC) for mosquito sampling in entomological studies on malaria transmission. Conflicting results have been obtained in comparing relative efficiency of alternative methods, according to the area, the species present and their density. The aim of this study was to compare the number and characteristics of mosquitoes sampled in two areas of Senegal by three different methods: HLC, light traps adjacent to an occupied bed net (LT/N), pyrethrum spray catches (PSC). Methods: Collections were performed in two villages: Dielmo (Soudan savanna) and Bandafassi (Soudan Guinean savanna), two or three nights per month for a 4-5 months period during the maximal transmission season in 2001-2002. Species were identified and Plasmodium infection determined by ELISA. The specific composition, circumsporozoite protein rate and entomological inoculation rate were calculated. Results: The diversity of mosquito species captured was maximal with LT/N, minimal with PSC. The mean number of anopheles captures each night was significantly different according to the method used and the species. PSC displayed a significantly lower anopheles density. HLC was the most efficient sampling method when Anopheles gambiae was the main vector (in Bandafassi); LT/N when it was Anopheles funestus (in Dielmo). A significant correlation was found between HLC and LT/M but correlation parameters were different according to the species. Circumsporozoite protein rates were not significantly different between methods or species. The entomological inoculation rate varied along with vector density and thus with methods and species. Conclusions: The choice of sampling method influenced entomological data recorded. Therefore, the sampling technique has to be chosen according to the vector studied and the aim of the study. Only HLC must be considered as the reference method, but in some conditions LT/N can be used as an alternative method.