Evaluating putative repellent ‘push’ and attractive ‘pull’ components for manipulating the odour orientation of host-seeking malaria vectors in the peri-domestic space

Background Novel malaria vector control approaches aim to combine tools for maximum protection. This study aimed to evaluate novel and re-evaluate existing putative repellent ‘push’ and attractive ‘pull’ components for manipulating the odour orientation of malaria vectors in the peri-domestic space. Methods Anopheles arabiensis outdoor human landing catches and trap comparisons were implemented in large semi-field systems to (i) test the efficacy of Citriodiol® or transfluthrin-treated fabric strips positioned in house eave gaps as push components for preventing bites; (ii) understand the efficacy of MB5-baited Suna-traps in attracting vectors in the presence of a human being; (iii) assess 2-butanone as a CO2 replacement for trapping; (iv) determine the protection provided by a full push-pull set up. The air concentrations of the chemical constituents of the push–pull set-up were quantified. Results Microencapsulated Citriodiol® eave strips did not provide outdoor protection against host-seeking An. arabiensis. Transfluthrin-treated strips reduced the odds of a mosquito landing on the human volunteer (OR 0.17; 95% CI 0.12–0.23). This impact was lower (OR 0.59; 95% CI 0.52–0.66) during the push-pull experiment, which was associated with low nighttime temperatures likely affecting the transfluthrin vaporisation. The MB5-baited Suna trap supplemented with CO2 attracted only a third of the released mosquitoes in the absence of a human being; however, with a human volunteer in the same system, the trap caught < 1% of all released mosquitoes. The volunteer consistently attracted over two-thirds of all mosquitoes released. This was the case in the absence (‘pull’ only) and in the presence of a spatial repellent (‘push-pull’), indicating that in its current configuration the tested ‘pull’ does not provide a valuable addition to a spatial repellent. The chemical 2-butanone was ineffective in replacing CO2. Transfluthrin was detectable in the air space but with a strong linear reduction in concentrations over 5 m from release. The MB5 constituent chemicals were only irregularly detected, potentially suggesting insufficient release and concentration in the air for attraction. Conclusion This step-by-step evaluation of the selected ‘push’ and ‘pull’ components led to a better understanding of their ability to affect host-seeking behaviours of the malaria vector An. arabiensis in the peri-domestic space and helps to gauge the impact such tools would have when used in the field for monitoring or control. Graphical Abstract