We show that our revised protocol for household aerosolized insecticides assessment using wipe-based decontamination, dual-cages, controlled-remote spray device and bioassay recording with action cameras are practical alternatives to enable higher throughput in the WHO current guidelines (15). For chamber decontamination, the 20 min wipe-based approach, is a major time-saving compared to the alternative (1 hour per chamber in our routine). Moreover, it also diminishes the length of user time within full protective equipment usage (e.g., respirator helmet and antistatic overalls). Furthermore, the approach minimises the likelihood of contaminating the room within which the PG chamber is housed as it produces less contaminated liquid and clothes for disposal compared to a full chamber’s washing. The wipe-based is also a worthwhile approach for semi-field testing room decontamination, in which a minimal furniture setup in distinct room layouts is applied (15, 16).
The wipe-based decontamination was effective for removing residual aerosolized insecticides, as evidenced by < 2% mortality of susceptible mosquitoes (< 20% threshold for unsatisfactory decontamination (15)) exposed to the chamber surface directly using cone tests (4). The wipe-based effectiveness could result from the scrubbing steps, which together with the 5% detergent solution act directly in the impregnated oil-based residuals on the chamber's surfaces. In contrast, after standard washing (detergent solution spraying, followed by a hosepipe rinse), we often recorded mortality higher than 20% for satisfactory decontamination, which spurred the investigation of alternatives. The detergent solution was also effective for removing pyrethroid-based residuals from cage fabric and metal frames (based on 2 hours of soaking), as indicated by zero mortality of confined susceptible mosquitoes in control cages after 24 hours.
A critical point addressed in our study and elsewhere (5, 17), is the challenge of standardizing exposure dose for ambient insecticides over testing procedures. As shown in Fig. 1B, despite a fixed burst length, the spray volume discharged varied across aerosol cans, which likely reflects manufacturing features (e.g., variable interior pressure, propellent volume, nozzle configuration, etc). Furthermore, the absence of a commercial remote spray device with the required features for research purposes (flexible spraying burst-length), imposes further challenges for inferring formulations dose-response on mosquitos' knockdown. Manual spraying through the chamber's door or access ports (where fitted) is likely to result in varying exposure doses. Indeed, our results highlight the need for, and importance of, reducing these sources of variation: for the A. aegypti resistant colony we observed a relationship between aerosol dose and whether susceptibility or resistance would be concluded (Fig. 2B). Ideally, the aerosol dose should be standardised by grams delivered rather than burst-length, facilitating comparison across studies.
To minimise the impact of technical variation, the RCAD is an alternative for future studies. Also, we are aware that normalising the density of an aerosol burst is not a feasible task, although based on our experience a calibration of the burst length can approximate the mass of aerosol delivered in a repeatable manner for different aerosol cans (Fig. 1B, 2A). Such standardisations to aerosol concentration delivery within any testing chamber are vital to improving the reproducibility of bioefficacy screening across mosquito strains and formulations.
Our dual-cage approach could facilitate aerosolized insecticide testing routines (e.g. tracking the evolution of insecticide resistance for household formulations) by boosting test capacity by 2- or 8-fold compared to the standard cage and free-flying assays, respectively. Our study design provided experimental evidence as shown in Fig. 4A, B, that the cage-based approach did not trigger heterogeneities in knockdown rate for insecticide resistance, prevent knockdown in the susceptible colonies, or delay knockdown (Fig. 3A-C).
Nevertheless, our results also revealed that the fan airflow is an important feature to improve the cage assay efficiency compared to the standard free-flight assays (Fig. 4C and Table 1). By contrast, the fan airflow has no impact on the free-flying mortality rate although was observed a gathering of mosquitoes knockdown around the fan for 1-hour operating (Fig. 4D). We observed that the knockdown gathering was linked to a vacuum effect instead of an airflow turbulence. It is important to bear in mind, that our results from free-flying are based on an assay set-up with an operating fan for aerosol dispersion, which is not applied by the current WHO guideline. However, a previous study testing vaporizing mats (17), also discussed the critical impact of airflow circulation on reproducibility and bias in result, as demonstrated by delayed and variable knockdown patterns underlined by the presence/absence of an operating fan and airflow direction.
In our assay set-up, fan level also impacted the bioassay results from different parts of the chamber; almost two-fold for the resistant Cayman strain (Fig. 3A). The solution to this issue is the simple, use of a spirit level to ensure a straight upward direction to avoid uneven aerosol dispersal. Similarly, a previous study also attributed inconsistencies in estimated mosquito coil efficacy between labs – and field-based testing due to limited ventilation and insecticide dispersion (5). Taken together, these insights also raise concerns about the absence of a formal guideline for semi-field study design, as often such studies (9, 16, 18) have adapted a test room without standardizing (e.g., room size, airflow circulation and manner of aerosol deployment), as well as using a cage-based assay without ventilation to disperse the insecticide.
In our experience and that of other research groups (5, 17), the recommendations for testing aerosol assays against free-flying mosquitoes in the current guideline (14), present substantial logistical challenges for laboratory and semi-field assessments. For instance, the free-flying approach has very low throughput (including the time-consuming recapture of free-flying mosquitoes) and presents technical drawbacks, such as the impaired ability to determine real-time knockdown effect over semi-field testing (17, 18).
While our validated dual-cage assay, which addresses such throughput limitations, is a feasible approach for screening mosquitos’ susceptibility against aerosolized insecticides, it is important to highlight that the recorded mortality for cage-assayed Cayman mosquitoes was significantly lower compared to free-flying (Fig. 4C, Table 1). The difference could reflect a reduced insecticide dose within the cages, as fewer aerosol droplets penetrate the mesh to the cage's interior. Further testing using fabrics with wider mash aperture, fan airflow strength and cage layouts such as cylindrical configurations (9, 16), could be performed to bring closer alignment between assay types.
Whether such quantitative differences observed in our present dual-cage assay set-up are important or not likely depends on the objective. When testing a new product, at least initially, the free flight assays may be preferred to allow the most precise assessment of quantitative effectiveness. However, if work involves the comparison between mosquito strains, the simultaneous testing in a more homogenized set-up presented by caged assays is likely to be advantageous, whereas a hybrid approach is also possible.
Remarkedly, despite recent reports of fuelled use of household insecticides for personal protection in vector-borne diseases endemic countries (7, 9, 19), likely due to the low availability of PG-chambers for research purposes and additional challenges in field-based testing, there is a dearth of published studies to permit insights into issues encountered when testing household insecticides. To this end, our experiential study provides suggestions to address limitations in the current guidelines (15) and to improve efficiency and throughput. Furthermore, the use of new features for the study design, like the use of action cameras in our routine, not only improved and refined the real-time data collection of the knockdown effect while screening up to eight strains in parallel but also would facilitate further studies of insect behaviour exposed to ambient insecticides. An additional footage file shows this in more detail (see Additional file 4).
The dual-cage assay set-up, which allows testing multiple strains in parallel, represents a time and cost-effective platform for the study of evolving insecticide resistance to household formulations in mosquitoes from vector-borne diseases endemic regions, which could also threaten the effectiveness of public health anti-mosquito programmes.