Efficacy of a spatial repellent for control of Aedes-borne virus transmission: A cluster-randomized trial in Iquitos, Peru

Abstract

Over half the world's population is at risk for viruses transmitted by Aedes mosquitoes, such as dengue and Zika. The primary vector, Aedes aegypti, thrives in urban environments. Despite decades of effort, cases and geographic range of Aedes-borne viruses (ABVs) continue to expand. Rigorously proven vector control interventions that measure protective efficacy against ABV diseases are limited to Wolbachia in a single trial in Indonesia and do not include any chemical intervention. Spatial repellents, a new option for efficient deployment, are designed to decrease human exposure to ABVs by releasing active ingredients into the air that disrupt mosquito-human contact. A parallel, cluster-randomized controlled trial was conducted in Iquitos, Peru, to quantify the impact of a transfluthrin-based spatial repellent on human ABV infection. From 2,907 households across 26 clusters (13 per arm), 1,578 participants were assessed for seroconversion (primary endpoint) by survival analysis. Incidence of acute disease was calculated among 16,683 participants (secondary endpoint). Adult mosquito collections were conducted to compare Ae. aegypti abundance, blood-fed rate, and parity status through mixed-effect difference-in-difference analyses. The spatial repellent significantly reduced ABV infection by 34.1% (one-sided 95% CI lower limit, 6.9%; one-sided P value = 0.0236, z = 1.98). Aedes aegypti abundance and blood-fed rates were significantly reduced by 28.6 (95% CI 24.1%, ∞); z = -9.11) and 12.4% (95% CI 4.2%, ∞); z = -2.43), respectively. Our trial provides conclusive statistical evidence from an appropriately powered, preplanned cluster-randomized controlled clinical trial of the impact of a chemical intervention, in this case a spatial repellent, to reduce the risk of ABV transmission compared to a placebo.

Keywords: Aedes aegypti; arbovirus vector; clinical trial; spatial repellent; vector control.

Conflict of interest statement

The authors declare no competing interest.

Figures

Fig. 1.

Location of 26 study clusters in Iquitos and Punchana Districts, Loreto Department, Iquitos, Peru. Each cluster consisted of approximately 140 households with an average distance of 523 m (range 280–879 m) between clusters.

Fig. 2.

Study timeline. (Top) Human blood sampling, disease surveillance, and entomological monitoring in relation to deployment of the SR intervention. (Bottom) Intervention rollout between Aug. and Dec. 2016 by cluster. Horizontal numbers correspond to cluster numbers shown in Fig. 1.

Fig. 3.

Allocation and follow-up of the longitudinal cohort population during three blood collection periods (baseline [B], first [F], and second/final [S]). The majority (62%) of participants provided samples at each collection period, whereas some only participated during year 1 (B-F) or year 2 (F-S). Participants with a single sample were lost to follow-up, and four individuals moved or had two houses located in the SR. Placebo clusters are shown as removed at the baseline period for clarity.

Fig. 4.

Kaplan-Meier curves of arbovirus infection for 13 SR and 13 placebo clusters in qualifying participants measured by seroconversion (primary endpoint) by cluster. (A) Hazard rates by individual cluster. (B) Aggregated hazard rate.

Fig. 5.

Mean densities of adult female Aedes aegypti collected per household survey in 13 SR and 13 placebo clusters by study month. Shaded areas represent the 95% CI around the mean.