**1. Introduction**

Vector control continues to offer effective prevention of mosquito-borne disease globally [1]. Insecticide-treated nets (ITNs) have been an extremely effective control measure [2] because they interrupt malaria transmission in two ways, by reducing mosquito blood-feeding and by killing a proportion of mosquitoes that contact the nets [3,4]. Since 2015, however, malaria control progress has stalled, with the COVID-19 pandemic in 2020 placing additional constraints on malaria control efforts. Despite this, ITNs remain the current cornerstone of global malaria control [1].

To date, all ITNs contain pyrethroid insecticides, which reduce the number of bites that individuals sleeping under them receive even if the nets become old and torn, because pyrethroids inhibit mosquito flight and feeding responses [5,6]. However, mass deployment of pyrethroid ITNs globally has led to widespread pyrethroid resistance with varying mechanisms [7] observed in eighty two countries. To sustain the malaria control gains attributed to ITNs and to assist in reducing malaria by at least 90% by 2030 [8], several ITNs with different insecticide classes in combination with pyrethroids have been developed. These so-called "dual-insecticide ITNs" afford non-neurotoxic modes of action with no cross-resistance (chlorfenapyr), reduced fecundity, and fertility (pyriproxyfen, PPF), or increased susceptibilities to pyrethroids (piperonyl butoxide, PBO) [9–14]. Randomized control trials have demonstrated greater malaria control using dual-active ITNs compared to pyrethroid nets in areas of high pyrethroid resistance, with pyrethroid combined with PBO [15,16], or chlorfenapyr [17]. Operational research has indicated an additional public health benefit of chlorfenapyr [18] and pyriproxyfen [18] in combination with pyrethroid compared to pyrethroid-only ITNs.

New ITN products must demonstrate their continued effectiveness for malaria control up to three years after deployment through biological efficacy testing against mosquito vectors [19,20]. The current World Health Organization (WHO) guidelines for ITN testing outline bioassays that were designed to evaluate pyrethroids with rapid neurotoxic action against exposed mosquitoes, i.e., rapid incapacitation (knockdown), reduction in bloodfeeding, and killing within 24 h post-exposure. Chlorfenapyr requires the mosquito to be metabolically or physiologically active (as it is when encountering ITNs during host-

seeking) to metabolize the parent molecule into the potent n-dealkylated form that elicits mosquito mortality [21]. Mosquitoes are more metabolically active at night, when flying, host-seeking or active during their typical circadian rhythms, for which the "free-flying" WHO tunnel test is a more appropriate bioassay [22,23].

The WHO tunnel test is widely used to assess the bio-efficacy of ITNs under laboratory conditions. Despite predicting a similar bio-efficacy of pyrethroid [5] and chlorfenapyr ITNs [22,23] to those measured in gold-standard experimental hut trials, the tunnel test has several limitations. Firstly, the animal baits (rabbit or guinea pig) used are non-preferred hosts for malaria mosquitoes, especially the highly anthropophilic Afrotropical vectors *Anopheles gambiae*, *Anopheles funestus*, and *Anopheles arabiensis* [24,25]. Moreover, the use of animals includes welfare concerns, and it is costly to ensure that animals are well maintained under veterinary supervision. Secondly, the bioassay is conducted overnight for 12–15 h. There is evidence that mosquitoes interact with treated netting within the first 30 min of release [26], thus prolonging exposure time could overestimate outcomes. Thirdly, the current tunnel test uses one hundred mosquitoes per replicate, which is expensive for insectaries to produce. Owing to the significance of blood-feeding in the life cycle of the malaria mosquito, as well as its importance for malaria transmission between human and mosquito hosts, it is an important component of vector control product testing. Different baits could be used to minimize the limitations of using humans or live animal hosts. Membrane feeders have been widely deployed for evaluating topical mosquito repellents [27], transmission-blocking drugs and vaccines [28], and endectocides [29], as well as for mosquito rearing [30–36]. Moreover, the use of an artificial membrane has several advantages, including no animal welfare or ethical concerns, reduced chance of accidental disease transmission, simple logistics, and reproducibility [33,35–37]. Given the significance of host kairomones in encouraging mosquito feeding, worn socks may be added to augmen<sup>t</sup> the attractiveness of the membrane to mosquitoes [38].

Multiple parameters including bait [25], mosquito density [39], and duration of exposure to ITNs [40] may affect the outcomes measured in tunnel tests. Therefore, the current paper compared alternative baits, exposure times, and lower mosquito densities against the current gold standard test (100 mosquitoes, animal bait, and 12 h exposure) as outlined in the WHO ITN evaluation guideline [20] in an attempt to simplify the tunnel test to make it cheaper with a higher throughput for the evaluation of large numbers of ITNs as needed for the bio-efficacy durability monitoring of chlorfenapyr ITNs that must be evaluated in "free-flying" bioassays [41].

#### **2. Materials and Methods**
