Infected Mosquitoes Have Altered Behavior to Repellents: A Systematic Review and Meta-analysis

Abstract

Abstract Here we conducted a systematic review and meta-analysis to reach a consensus on whether infected and uninfected mosquitoes respond differently to repellents. After screening 2,316 published studies, theses, and conference abstracts, we identified 18 studies that tested whether infection status modulated the effectiveness of repellents. Thirteen of these studies had outcomes available for meta-analysis, and overall, seven repellents were tested (typically DEET with 62% of outcomes), six mosquito species had repellence behaviors measured (typically Aedes aegypti (L.) (Diptera: Culicidae) mosquitoes with 71% of outcomes), and a broad diversity of infections were tested including Sindbis virus (Togaviridae: Alphavirus) (33% of outcomes), Dengue (Flaviviridae: Flavivirus) (31%), malaria (Plasmodium berghei Vincke & Lips (Haemospororida: Plasmodiidae) or P. falciparum Welch (Haemospororida: Plasmodiidae); 25%), Zika (Flaviviridae: Flavivirus) (7%), and microsporidia (4%). Pooling all outcomes with meta-analysis, we found that repellents were less effective against infected mosquitoes—marking an average 62% reduction in protective efficacy relative to uninfected mosquitoes (pooled odds ratio = 0.38, 95% confidence interval = 0.22–0.66; k = 96). Older infected mosquitoes were also more likely to show altered responses and loss of sensitivity to repellents, emphasizing the challenge of distinguishing between age or incubation period effects. Plasmodium- or Dengue-infected mosquitoes also did not show altered responses to repellents; however, Dengue–mosquito systems used inoculation practices that can introduce variability in repellency responses. Given our findings that repellents offer less protection against infected mosquitoes and that these vectors are the most dangerous in terms of disease transmission, then trials on repellent effectiveness should incorporate infected mosquitoes to improve predictability in blocking vector–human contact.