Abstract
BACKGROUND: Aedes aegypti and Culex quinquefasciatus are the main urban vectors of arthropod-borne viruses causing human disease, including dengue, Zika, or West Nile. Although key to disease prevention, urban-mosquito control has met only limited success. Alternative vector-control tactics are therefore being developed and tested, often using entomological endpoints to measure impact. Here, we test one promising alternative and assess how three such endpoints perform at measuring its effects. METHODS: We conducted a 16-month, two-arm, cluster-randomized controlled trial (CRCT) of mosquito-disseminated pyriproxyfen (MD-PPF) in central-western Brazil. We used three entomological endpoints: adult-mosquito density as directly measured by active aspiration of adult mosquitoes, and egg-trap-based indices of female Aedes presence (proportion of positive egg-traps) and possibly abundance (number of eggs per egg-trap). Using generalized linear mixed models, we estimated MD-PPF effects on these endpoints while accounting for the non-independence of repeated observations and for intervention-unrelated sources of spatial-temporal variation. RESULTS: On average, MD-PPF reduced adult-mosquito density by 66.3% (95% confidence interval, 95% CI: 47.3-78.4%); Cx. quinquefasciatus density fell by 55.5% (95% CI: 21.1-74.8%), and Ae. aegypti density by 60.0% (95% CI: 28.7-77.5%). In contrast, MD-PPF had no measurable effect on either Aedes egg counts or egg-trap positivity, both of which decreased somewhat in the intervention cluster but also in the control cluster. Egg-trap data, therefore, failed to reflect the 60.0% mean reduction of adult Aedes density associated with MD-PPF deployment. CONCLUSIONS: Our results suggest that the widely used egg-trap-based monitoring may poorly measure the impact of Aedes control; even if more costly, direct monitoring of the adult mosquito population is likely to provide a much more realistic and informative picture of intervention effects. In our CRCT, MD-PPF reduced adult-mosquito density by 66.3% in a medium-sized, spatially non-isolated, tropical urban neighborhood. Broader-scale trials will be necessary to measure MD-PPF impact on arboviral-disease transmission.