The relationship between entomological indicators of Aedes aegypti abundance and dengue virus infection

Abstract

Routine entomological monitoring data are used to quantify the abundance of Ae. aegypti. The public health utility of these indicators is based on the assumption that greater mosquito abundance increases the risk of human DENV transmission, and therefore reducing exposure to the vector decreases incidence of infection. Entomological survey data from two longitudinal cohort studies in Iquitos, Peru, linked with 8,153 paired serological samples taken approximately six months apart were analyzed. Indicators of Ae. aegypti density were calculated from cross-sectional and longitudinal entomological data collected over a 12-month period for larval, pupal and adult Ae. aegypti. Log binomial models were used to estimate risk ratios (RR) to measure the association between Ae. aegypti abundance and the six-month risk of DENV seroconversion. RRs estimated using cross-sectional entomological data were compared to RRs estimated using longitudinal data. Higher cross-sectional Ae. aegypti densities were not associated with an increased risk of DENV seroconversion. Use of longitudinal entomological data resulted in RRs ranging from 1.01 (95% CI: 1.01, 1.02) to 1.30 (95% CI: 1.17, 1.46) for adult stage density estimates and RRs ranging from 1.21 (95% CI: 1.07, 1.37) to 1.75 (95% CI: 1.23, 2.5) for categorical immature indices. Ae. aegypti densities calculated from longitudinal entomological data were associated with DENV seroconversion, whereas those measured cross-sectionally were not. Ae. aegypti indicators calculated from cross-sectional surveillance, as is common practice, have limited public health utility in detecting areas or populations at high risk of DENV infection.

Author summary

In this study, we compared measures of entomological risk collected through routine household entomological monitoring by estimating an association with human DENV infection. Longitudinal entomological and human serology data from Iquitos, Peru, were used to test associations between Ae. aegypti indices and the 6-month risk of DENV seroconversion. Our analysis found no association between cross-sectional measures of Ae. aegypti abundance and the risk of DENV seroconversion. Longitudinal measures of Ae. aegypti were better proxies for DENV risk, primarily among adult stage mosquito indicators. DENV transmission is complex and time-varying; the relationship between vector density and risk is not static nor adequately characterized through periodic entomological surveillance. While entomological monitoring will continue to serve a role in the evaluation of vector control interventions (e.g., comparing pre- and post-intervention abundance), our analysis challenges the validity of most Ae. aegypti indicators as adequate proxies for true DENV exposure risk.