Malaria vectors diversity, insecticide resistance and transmission during the rainy season in peri-urban villages of south-western Burkina Faso

The survey was carried out during the 2015 rainy season in the—29 km apart—peri-urban areas of Diébougou (N10.96741; W 003.24580) and Dano (N11.14288; W 003.05969) cities, in South-West Burkina Faso (Fig. 1). Both areas have similar environmental characteristics with an average 1000 mm annual rainfall occurring from May to October and a vegetation dominated by wooded savannah. Surveys were performed in height sites named Diébougou centre, Bagane, Loto and Bapla in the Diébougou area; and sector one to four in Dano. The main economic activity is agriculture in both areas where animals and humans use to live very closely in the same courtyard. In the Diébougou area, bendiocarb IRS was implemented in 2011 and 2012. Populations of both areas received free of charge LLINs in 2010 and 2013 in the framework of the NMCP national mass distribution.

Monthly mosquito collections were carried out from August to November 2015. One inhabited house was randomly selected in each site (i.e. 4 houses per study area). These houses were made of mud or cement with traditional roof or metal sheeting and were representative of the local housing. Mosquito collections were performed using CDC light trap both indoor and outdoor of each selected house during four successive nights from 18:00 to 06:00. Thus, 32 trap-night collections were performed per month and per area. Each month, early morning mosquito collections were also performed using pyrethrum spraying catches (PSC) inside 10 randomly selected inhabited houses over 4 consecutive days in each site (for a total of 40 houses per area per month).

All collected Anopheles adults were morphologically identified under a stereomicroscope to the species complex and preserved in 1.5 ml tubes containing silicagel [20]. Unfed females form CDC light traps belonging to the An. gambiae complex were dissected to determined their physiological age using the Detinova method [21]. A randomly selected sub-sample (25 female per house per month) of females collected in light traps and belonging to the An. gambiae complex were proceed by Polymerase Chain Reaction (PCR) for species identification following the protocol described by Santolamazza et al. [22]. Female infection by Plasmodium falciparum was also assessed on the same subsample using enzyme-linked immunosorbent assays (ELISA) technique described by Wirtz et al. [23]. Abdomen of blood-fed females collected indoor by PSC were analysed for identification of blood meal source by the ELISA technique [24].

Anopheles larvae were collected from natural breeding sites throughout the two peri-urban areas of Diébougou and Dano. In each study area, mosquito larvae were collected in at least 10 breeding sites distant from each other by at least 200 m. Larvae were then pooled together, brought back to the IRSS insectarium and reared under controlled conditions (temperature 27 ± 2 °C, Relative humidity 70 ± 10%) until adult’s emergence. Non-blood-fed 3–5 days-old females morphologically identified as An. gambiae s.l. were put in contact with DDT 4%, deltamethrin 0.05%, bendiocarb 0.1% and chlorpyriphos methyl 0.1% impregnated filter papers following the WHO standard protocol [25]. Four replicates of 20–25 individuals were exposed to each tested insecticide. Anopheles gambiae “Kisumu” strain was used as the susceptible control strain [25]. Mortality was recorded 24 hours after exposure. PCR analyses were then conducted on a subsample of 200 females per area to detect the kdr (L1014F) and ace-1 (G119S) mutations using the protocol described by Martinez-Torres et al. [26] and Weill et al. [27], respectively.

Mosquito density per trap per night was calculated for each malaria vector species as the number of Anopheles individuals collected per trap per night. Endophagy rate (ER) was the proportion of mosquito collected indoors from CDC light traps. Parous rate (PR) was calculated as the proportion of parous An. gambiae s.l. among dissected individuals. The Plasmodium- sporozoite rate (SR) of infection was calculated as the proportion of mosquitoes positive for CSP-ELISA. Human Blood Index (HBI) was the proportion of mosquitoes found to be fed on Humans relative to the total number tested.

All other statistical analyses were performed using the software R version 3.6.0 [28]. A generalized linear mixed model (GLMM) fitting a negative binomial distribution of the error was applied to compare Anopheles densities between areas and species. The collection site, position (indoor or outdoor) and date were included in the model as random intercepts (sites and positions were nested). The human blood index of both An. gambiae s.l. and Anopheles funestus were compared between areas using logistic regression. The SR, PR, and ER were compared between areas using binomial GLMMs fitted on individual data and for both An. gambiae s.l. and An. funestus. The collection site and date were included in the model as random intercepts. GLMM were fitted using the glmmTMB function of the glmmTMB package [29]. The post-hoc Tukey’s method was used to perform multiple comparison among modalities of the fixed terms. The ‘emmeans’ function of the ‘emmeans’ package [30] was used to compute Density rate ratios (DRR) and Odds ratios (OR) with 95% confidence intervals for Negative Binomial and Binomial models, respectively. The allelic frequencies of the kdr and ace-1 mutations in An. gambiae s.l. populations were calculated and compared using the ‘GenePop’ package in R [31].