Insecticide-treated bednets (ITNs) are being strongly promoted as a malaria control tool in Africa by the World Health Organization and other international agencies 1. Their efficacy in reducing man-vector contact, malaria morbidity and mortality has been demonstrated in various epidemiological situations 2345. With current use of pyrethroids in agriculture and increasing scale of ITNs coverage, selective pressure for pyrethroid resistance in mosquitoes is expected to increase 67. Resistance to pyrethroids has been reported in both Anopheles gambiae and Anopheles funestus in many malaria endemic countries in Africa 89101112, prompting for alternative products and strategies for malaria vector control 1314. Studies in experimental huts and village scale trials however, suggest that ITNs maintain their efficacy in areas with high frequency of kdr and insensitive acetylcholinesterase (Ace.I^R) resistance genes in An. gambiae 1516171819. This should be investigated in areas where pyrethroid-resistance mechanisms other than these well-known target site mutations are incriminated 2021. Indeed, enzymes involved in insecticide detoxification may further jeopardize malaria vector control with insecticide-treated materials. This was recently exemplified in South Africa, where metabolic resistance to pyrethroids in An. funestus required a switch back from pyrethroids to DTT for house spraying to restore malaria control 9.

In this study, the efficacy of bednets impregnated with bifenthrin against natural populations of An. gambiae s.l. showing metabolic-based resistance was assessed in experimental huts. Artificially holed bednets were used in an attempt to mimic the damage that commonly occurs through domestic use. The effect of this treatment was assessed on the local mosquito populations, and the level of personal and mass protection expected from such control strategy was estimated.

Insecticide Treated Nets (ITNs) not only provide personal protection; a high level of coverage in a population may benefit every individual in the community by contributing to an area-wide reduction in malaria vector populations 19. The personal and mass effect protection that could be conferred by bifenthrin was estimated according to WHO guidelines 29. The personal protection effect of a treatment relative to the control was estimated by the formula: 100 × (B_u-B_t)/B_u, where B_u is the total number of blood-fed mosquitoes in the untreated hut, and B_t is the total number of blood-fed mosquitoes in the treated hut. The overall insecticidal effect of a treatment was estimated by the formula: 100 × (D_t-D_u)/E_u where D_t is the total number of mosquitoes dying in the treated hut, D_u is the number dying in the untreated hut, and E_u is the number of mosquitoes entering the untreated hut 29. For both species, personal protection conferred by bifenthrin at 50 mg/m² was very high (more than 90%) while it was inexistent at 5 mg/m². Conversely, the mass effect protection was very low, particularly at 50 mg/m² (Table 2).

The high level of personal protection conferred by bifenthrin at 50 mg/m² is a reflection of the significant level of blood feeding inhibition combined with the strong deterrent effect observed on the Pitoa mosquito populations (Table 1). Recent studies using the tunnel test technique in controlled condition also attributed this high inhibition of blood feeding to the irritant effect of bifenthrin 30, although it was found to be less irritant than other pyrethroids in laboratory conditions 2324. Hence, although the mortality rate was high with bifenthrin at 50 mg/m², the deterrent effect greatly reduced the number of mosquitoes killed, thus resulting in a low mass protection effect.

Therefore, the deterrent effect, i.e. the reduction in the number of mosquitoes entering the hut, is not a reliable indicator of ITNs efficacy. Moreover, the values of this index vary considerably between experiments, e.g. from zero to 70.0% against kdr-resistant An. gambiae, even at the same study site in Côte d'Ivoire 2526. As a result, and acknowledging the absence of comparative studies involving enzyme-based pyrethroid resistance mechanisms, the high reduction of entry rate induced by bifenthrin in this study remains tentative and difficult to interpret. However, an obvious dose dependency of the deterrent effect was noted on both An. gambiae and An. funestus (Table 1).

For the same reason, the induced exophily does not permit an easy interpretation of the results. While exophily induced by bifenthrin is low in Côte d'Ivoire against An. gambiae (maximum value: 24.0%) 25, it was higher (Table 1) against the resistant An. gambiae population from Pitoa but less so against An. funestus. This is consistent with a former report demonstrating that An. funestus was 3 times less irritable as An. gambiae to DDT in North Cameroon 31. It is tempting to speculate that such variability in the sensitivity of these vector species to the excito-repellent effect of insecticides applies with bifenthrin, a type I pyrethroid with the same neuro-physiological mode of action than DDT. Furthermore, the great majority of An. gambiae s.l. identified in Pitoa was constituted by An. arabiensis, which might be differentially affected by the excito-repellent effect of bifenthrin than its sibling species. This might reflect species-specific variability in behavioural traits linked to host-seeking, biting and/or feeding behaviour and suggests that mosquito behaviour might modulate to a considerable extent the efficacy of insecticides and ITNs in natural conditions.

Almost all the An. arabiensis and An. funestus fed in the presence of the untreated net but high dose of bifenthrin markedly reduced these rates in both vectors (Table 1). In addition, a great proportion of An. arabiensis and An. funestus entering the hut with a net treated at 50 mg/m² were found dead at the time of collection, in similar proportions to what has been reported with bifenthrin 2526, but in higher proportions than what was observed with deltamethrin 16. As such, treatment with the high dose bifenthrin restored the protectiveness of torn nets against An. arabiensis and An. funestus, as has been demonstrated with various pyrethroid treatments and against various mosquito species and populations 1932333435. Both blood-feeding inhibition and mortality rates appeared to be dosage-dependant, inasmuch as the effect of bifenthrin at 5 mg/m² on these indicators was low and, in most case, non-significant (Table 1). Further tests with intermediate concentration would however be needed to confirm a dosage trend.

Moreover, no data are currently available on the susceptibility of An. funestus to insecticides in this area and this needs to be monitored, acknowledging enzyme-based insecticide resistance was recently detected in populations from South Africa 9. Additionally, although resistance to permethrin and deltamethrin was demonstrated in An. gambiae s.l. from Pitoa 12, susceptibility to bifenthrin has not yet been evaluated and cross-resistance between these compounds should not be assumed due to high substrate specificity and selectivity of the enzymes involved in pyrethroids detoxification pathways 3637.

The results presented above confirm that bifenthrin at 50 mg/m² is a suitable dosage that could be recommended for mosquito net impregnation, even in areas of vector resistance to commonly used pyrethroids (permethrin and deltamethrin) and regardless of the mechanisms involved (kdr or enzyme-based). Bifenthrin impregnated bed nets offer considerable personal protection, even in the presence of holes in the net. However, its strong deterrent effect might jeopardize any area-wide impact on vector populations, as should be expected from large scale intervention programs.

MC carried out the field evaluation of nets, analysed and interpreted data and drafted the manuscript. FS participated in the conception of the study and coordinated its implementation in the fields, and substantially helped draft the manuscript and data interpretation. FC participated in the conception of the study and its design and helped in data analysis and interpretation. JE participated in data analysis and interpretation. FD set up the field station in Pitoa & JMH planned the study and its design, participated in data analysis and interpretation, and helped draft the manuscript.