Temporal correlations between mosquito-based dengue virus surveillance measures or indoor mosquito abundance and dengue case numbers in Mérida City, México.

Surveillance of dengue virus (DENV) in Aedes (Stegomyia) aegypti (L.) females is of potential interest because human DENV infections are commonly asymptomatic, which decreases the effectiveness of dengue case surveillance to provide early warning of building outbreaks. Our primary aim was to examine if mosquito-based virological measures--monthly percentages of examined Ae. aegypti females infected with DENV or examined homes from which at least one DENV-infected Ae. aegypti female was collected--are correlated with reported dengue cases in the same or subsequent months within study neighborhoods in Méida City, México. The study encompassed approximately 30 neighborhoods in the southern and eastern parts of the city. Mosquitoes were collected monthly over a 15-mo period within study homes (average of 145 homes examined per month); this produced approximately 5,800 Ae. aegypti females subsequently examined for DENV RNA. Although monthly dengue case numbers in the study neighborhoods varied > 100-fold during the study period, we did not find statistically significant positive correlations between monthly data for mosquito-based DENV surveillance measures and reported dengue cases in the same or subsequent months. Monthly average temperature, rainfall, and indoor abundance of Ae. aegypti females were positively correlated (P < or = 0.001) with dengue case numbers in subsequent months with lag times of 3-5, 2, and 1-2 mo, respectively. However, because dengue outbreak risk is strongly influenced by serotype-specific susceptibility of the human population to DENV, the value of weather conditions and entomological indices to predict outbreaks is very limited. Potential ways to improve the sensitivity of mosquito-based DENV surveillance are discussed.