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

Elizabeth A Cromwell,William B Messer,Christopher M Barker,Steven R Meshnick,Amy C Morrison,Annelies Van Rie,Steven T Stoddard,Thomas W Scott,Duane J Gubler

doi:10.1371/journal.pntd.0005429

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.

Highlights

Dengue virus (DENV), which is transmitted by the bite of female Aedes aegypti mosquitoes, causes more human morbidity and mortality than any other arthropod-borne virus [1]
The principal finding of this analysis is that a higher household level Ae. aegypti density calculated from cross-sectional entomological data was not associated with an increase in the risk of DENV infection
By comparing measures of Ae. aegypti density calculated from cross-sectional data to an average density, we are able to explore the potential for non-differential measurement error of mosquito abundance to bias the association between Ae. aegypti monitoring indicators and DENV infection towards the null

Summary

Introduction

Dengue virus (DENV), which is transmitted by the bite of female Aedes aegypti mosquitoes, causes more human morbidity and mortality than any other arthropod-borne virus [1]. The World Health Organization recommends monitoring vector abundance for the targeting and evaluation of vector control interventions [5]. The frequency of entomological data collection varies by setting, and WHO guidelines recommend implementation occur at a frequency from “weeks to months” [5]. Monitoring indicators vary by mosquito life stage (adults, larvae and/or pupae), availability of larval development sites (infestation indices), and process of collection (fixed trap or human-based surveys such as adult aspirator collections, household inspection for larvae) [9]. The public health utility of these indicators is based on the assumption that greater mosquito abundance increases the risk of DENV transmission, and reducing exposure to the vector decreases incidence of infection. By identifying “hot spots” of Ae. aegypti infestation, targeted vector control would be an efficient use of limited intervention resources [10]

Objectives

Methods

Results

Discussion

Conclusion