Assessing the effects of interventions for Aedes aegypti control: systematic review and meta-analysis of cluster randomised controlled trials

Abstract Background The Aedes aegypti mosquito is the vector for dengue fever, yellow fever, chikungunya, and zika viruses. Inadequate vector control has contributed to persistence and increase of these diseases. This review assesses the evidence of effectiveness of different control measures in reducing Aedes aegypti proliferation, using standard entomological indices. Methods A systematic search of Medline, Ovid, BVS, LILACS, ARTEMISA, IMBIOMED and MEDIGRAPHIC databases identified cluster randomised controlled trials (CRCTs) of interventions to control Aedes aegypti published between January 2003 and October 2016. Eligible studies were CRCTs of chemical or biological control measures, or community mobilization, with entomological indices as an endpoint. A meta-analysis of eligible studies, using a random effects model, assessed the impact on household index (HI), container index (CI), and Breteau index (BI). Results From 848 papers identified by the search, eighteen met the inclusion criteria: eight for chemical control, one for biological control and nine for community mobilisation. Seven of the nine CRCTs of community mobilisation reported significantly lower entomological indices in intervention than control clusters; findings from the eight CRCTs of chemical control were more mixed. The CRCT of biological control reported a significant impact on the pupae per person index only. Ten papers provided enough detail for meta-analysis. Community mobilisation (four studies) was consistently effective, with an overall intervention effectiveness estimate of −0.10 (95%CI -0.20 – 0.00) for HI, −0.03 (95%CI -0.05 – -0.01) for CI, and −0.13 (95%CI -0.22 – -0.05) for BI. The single CRCT of biological control had effectiveness of −0.02 (95%CI -0.07– 0.03) for HI, −0.02 (95%CI -0.04– -0.01) for CI and −0.08 (95%CI -0.15– -0.01) for BI. The five studies of chemical control did not show a significant impact on indices: the overall effectiveness was −0.01 (95%CI -0.05– 0.03) for HI, 0.01 (95% CI -0.01– 0.02) for CI, and 0.01 (95%CI -0.03 – 0.05) for BI. Conclusion Governments that rely on chemical control of Aedes aegypti should consider adding community mobilization to their prevention efforts. More well-conducted CRCTs of complex interventions, including those with biological control, are needed to provide evidence of real life impact. Trials of all interventions should measure impact on dengue risk.