Background
Outbreaks of dengue fever caused by viruses transmitted by Aedes aegypti mosquitoes are repeated occurrences in West Africa. In recent years, Burkina Faso has experienced major dengue outbreaks, most notably in 2016 and 2017 when 80% of cases were recorded in Ouagadougou City (Central Health Region). In order to better understand the ecology of this vector and to provide information for use in developing control measures, a study on the characteristics of Aedes container breeding sites and the productivity of such sites, as measured by the abundance of immature stages and resultant adult body size, was undertaken in three health districts (Baskuy, Bogodogo and Nongremassom) of Ouagadougou.

Methods
Adult mosquitoes were collected indoors and outdoors in 643 households during the rainy season from August to October 2018. The presence of water containers was systematically recorded and the containers examined for the presence or absence of larvae. Characteristics of the container breeding sites, including size of the container and temperature, pH and conductivity of the water contained within, were recorded as well as the volume of water. Traditional Stegomyia indices were calculated as quantitative indicators of the risk of dengue outbreaks; generalised mixed models were fitted to larval and pupal densities, and the contribution of each covariate to the model was evaluated by the Z-value and associated P-value.

Results
A total of 1061 container breeding sites were inspected, of which 760 contained immature stages of Ae. aegypti (‘positive’ containers). The most frequent container breeding sites found in each health district were tyres and both medium (buckets/cans/pots) and large (bins/barrels/drums) containers; these containers were also the most productive larval habitats and the types that most frequently tested positive. Of the Stegomyia indices, the Breteau, House and Container indices exceeded WHO dengue risk thresholds. Generalised linear mixed models showed that larval and pupal abundances were associated with container type, physicochemical characteristics of the water and collection month, but there were significant differences among container types and among health districts. Aedes aegypti body size was positively associated with type and diameter of the container, as well as with electrical conductivity of the water, and negatively associated with pH and temperature of the water and with the level of exposure of the container to sunlight.

Conclusion
This study provides data on putative determinants of the productivity of habitats regarding Ae. aegypti immature stages. These data are useful to better understand Ae. aegypti proliferation. The results suggest that identifying and targeting the most productive container breeding sites could contribute to dengue vector control strategies in Burkina Faso.

Ae. aegypti, Breeding sites, Larvae, Pupae, Stegomyia index, Body size, Dengue