Using adult Aedes aegypti females to predict areas at risk for dengue transmission: A spatial case-control study

Abstract

Introduction: Traditional indices for measuring dengue fever risk in a given area are based on the immature forms of the vector (larvae and pupae surveys). However, this is inefficient because only adult female mosquitoes actually transmit the virus. Based on these assumptions, our objective was to evaluate the association between an entomological index obtained from adult mosquito traps and the occurrence of dengue in a hyperendemic area. Additionally, we compared its cost to that of the Breteau Index (BI). Material and methods: We performed this study in São José do Rio Preto, SP, Brazil, between the epidemiological weeks of 36/2012 and 19/2013. BG-Sentinel and BG-Mosquitito traps were installed to capture adult mosquitoes. Positive and negative cases of dengue fever were computed and geocoded. We generated biweekly thematic maps of the entomological index, generated by calculating the number of adult Aedes aegypti females (NAF) per 100 households during a week by kriging, and based on the number of mosquitoes captured. The relation between the occurrence of dengue fever and the NAF was tested using a spatial case-control design and a generalized additive model and was controlled by the coordinates of the positive and negative cases of dengue fever. Results: Our analyses showed that increases in dengue fever cases occurred in parallel with increases in the number of Ae. aegypti females. The entomological index produced in our study correlates positively with the incidence of dengue, particularly during intervals when vector control measures were applied less intensively. The operational costs of our index were lower than those of the BI: NAF used 71.5% less human resources necessary to measure the BI. Conclusions: Spatial analysis techniques and the number of adult Ae. aegypti females were used to produce an indicator of dengue risk. The index can be applied at various levels of spatial aggregation for an entire study area, as well as for sub-areas, such as city blocks. Even though the index is adequate to predict dengue risk, it should be tested and validated in various scenarios before routine use.