Modelling spatiotemporal trends in the frequency of genetic mutations conferring insecticide target-site resistance in African mosquito malaria vector species

Background Resistance in malaria vectors to pyrethroids, the most widely used class of insecticides for malaria vector control, threatens the continued efficacy of vector control tools. Target-site resistance is an important genetic resistance mechanism caused by mutations in the voltage-gated sodium channel (Vgsc) gene that encodes the pyrethroid target-site. Understanding the geographic distribution of target-site resistance, and temporal trends across different vector species, can inform strategic deployment of vector control tools. Results We develop a Bayesian statistical spatiotemporal model to interpret species-specific trends in the frequency of the most common resistance mutations, Vgsc-995S and Vgsc-995F, in three major malaria vector species Anopheles gambiae, An. coluzzii, and An. arabiensis over the period 2005–2017. The models are informed by 2418 observations of the frequency of each mutation in field sampled mosquitoes collected from 27 countries spanning western and eastern regions of Africa. For nine selected countries, we develop annual predictive maps which reveal geographically structured patterns of spread of each mutation at regional and continental scales. The results show associations, as well as stark differences, in spread dynamics of the two mutations across the three vector species. The coverage of ITNs was an influential predictor of Vgsc allele frequencies, with modelled relationships between ITN coverage and allele frequencies varying across species and geographic regions. We found that our mapped Vgsc allele frequencies are a significant partial predictor of phenotypic resistance to the pyrethroid deltamethrin in An. gambiae complex populations. Conclusions Our predictive maps show how spatiotemporal trends in insecticide target-site resistance mechanisms in African An. gambiae vary across individual vector species and geographic regions. Molecular surveillance of resistance mechanisms will help to predict resistance phenotypes and track their spread.