Our study shows a widespread resistance of An. gambiae to deltamethrin in the Plateau region, southeastern Benin. Taking into account the KdTs, the effect of deltamethrin on An. gambiae mosquitoes was not uniform across the various localities. Resistance with deltamethrin was associated with relatively high kdr frequencies found in both M and S form of An. gambiae s.s.
Out of the 56 localities explored, An. gambiae was found susceptible in only two localities (Ko-Aidjèdo and Zihan). These results show a wide distribution of deltamethrin resistance as previously reported by Padonou et al. in certain localities of Ouéme, a department near Plateau region. The results also confirm the rapid expansion of pyrethroid resistance in natural populations of An. gambiae in Africa  and particularly in Benin [21–23]. However, the small numbers of mosquitoes exposed to deltamethrin in some localities did not allow accurate estimations of the mortality rates. This constitutes a major limitation for our study.
Previous studies demonstrated that malaria vector resistance to pyrethroids might be related to an extensive and massive use of LLINs [24, 25]. Insecticide molecules at the surface of LLINs might exert a lethal effect on susceptible mosquitoes, therefore selecting for resistant mosquitoes that reproduce in natural populations. In the department of Ouémé for instance, Padonou et al.  reported an increase in the resistance of An. gambiae to deltamethrin following a mass distribution campaign of LLINs. It is then possible that the resistance level has increased very quickly in the region of Plateau after the selective distribution campaign of LLINs in 2007. Several other factors such as the domestic use of pyrethroids  and water run off loaded with insecticide particles from the North of Benin where pyrethroids were massively and uncontrollably used to control cotton pests  could also explained the resistance of An. gambiae to deltamethrin in the localities of the region of Plateau. The insecticides particles contained in the water could exert a selection pressure on the larvae of An. gambiae.
Our data surprisingly reveals a low resistance of An. gambiae to deltamethrin in the district of Ifangni. Nonetheless, we did not observe any spatial variation in the level of resistance across the four districts (p = 0.22). This suggests a similar selection pressure on An. gambiae s.l across the four districts. Therefore, other reasons, still unknown to our knowledge, may explain the low levels of resistance of Anopheles population to deltamethrin in the district of Ifangni. Further studies must be undertaken to understand the real causes of the low resistance of An. gambiae in this district.
PCR analysis reveals that An. gambiae s.s was the only species of the An. gambiae s.l. complex encountered in all 16 selected localities. The absence of An. melas could be explained by the fact that the larvae were all collected from small freshwater pools. This is understandable since An. melas larvae are mainly found in brackish water ponds . Similarly, An. arabiensis was not present even though it has already been reported in Central Benin .
Regarding the molecular forms, the relative dominance of one form over the other could be explained by the presence of specific breeding sites to one or the other molecular forms . Globally, in both areas (R+ and R+++), 66.8% of the mosquitoes were of M form, 32.6% of S form and 0.6% of the hybrid M/S. This contradicts Yadouléton et al. who reported the absence of the form S of An. gambiae s.s in the region of Plateau. In fact the presence of the form S in significant proportions in 2011 in that region could be associated to the rainfall pattern and the rapid infiltration of water into the soil allowing the formation of favorable temporary breeding sites for the development of form S of An. gambiae s.s.
Kdr mutation 1014 F has been found in M as well S form of An. gambiae s.s, but at variable frequencies depending on the locality. Djènontin et al. reported a higher frequency of kdr mutation in the form S than in the form M of An. gambiae s.s in Ouidah-Kpomassè-Tori Bossito area between October and December 2007. The same trend has been reported by Diabaté et al. and Dabiré et al. in the Kou valley in Burkina-Faso and by Dabiré et al. in Guinea-Bissau. However, for our study, data from both areas (R+ and R+++) revealed a similar kdr frequency between the two molecular forms (p = 0.17). This suggests that the selection pressure exerted on the two molecular forms in the natural environment did not differ significantly across the region.
We did not find the 1014S kdr mutation originally from East Africa . However, previous studies have indicated that this mutation was present in Benin  and was expanding to the north and center of Benin in An. gambiae. It is therefore important to extend the surveillance of kdr1014S in the region of Plateau.
Although a significant difference in mortality rates was found between both areas (R+ and R+++) (p < 0.0001), no difference was observed between the frequencies of kdr gene in both areas (p = 0.13). In addition, KdT95 was higher in R+++ area than in R+ area (p < 0.05). These findings suggest the involvement of other resistance mechanisms in addition to the kdr mutation in the mosquitoes from the R+++ area.
Recent studies have also reported the involvement of certain metabolic enzymes in the resistance of malaria vectors to pyrethroids in Africa  and in several other regions in Benin [36, 37]. For example, an overexpression of CYP6M2 and CYP6P3 genes, involved in the metabolism of pyrethroids, has been reported in resistant populations of An. gambiae in Porto Novo . Therefore, we believe that the role of metabolic resistance in the region of Plateau deserves further scrutiny.
From the results recorded from the susceptibility tests, we were able to map the distribution of the resistance of An. gambiae to deltamethrin. Such a map was necessary since it provides a picture of the availability of localities where An. gambiae was still susceptible to deltamethrin. In addition, this map could be used as an important tool to monitor the dynamics of the resistance of An. gambiae to pyrethroids. Moreover, given that our research activities were carried out in collaboration with Benin NMCP, this map should allow Benin NMCP to accordingly adapt its strategy of malaria prevention.
Given the extensive use of LLINs inside houses at this time, it is possible that highly resistant mosquitoes would tend to feed inside more than low resistant Anopheles mosquitoes. The deterrence effect of LLINs is likely to be more effective on low resistant Anopheles population. If this was verified, the likelihood for highly resistant mosquitoes to transmit malaria parasites would be higher than that of lowly resistant ones. If this assumption was true, it would confirm that vector resistance to insecticides is a major concern to the operational effectiveness of the LLINs distributed in 2011 .