The results of LLIN effectiveness in malaria prevention in vector resistance area showed that the resistance of vectors does not reduce the effectiveness of LLINs, but the prevalence of malaria and anaemia was higher in low resistance areas, and was in contradiction with what was expected. Anopheles gambiae, the main vector of malaria in Africa, has developed a strong resistance to pyrethroid in southern Benin . This resistance has been observed not only in urban areas and in areas characterized by cotton growing but also in rural areas where traditional farming does not require the use of agricultural insecticides or fertilizers [11, 15, 26]. The main mechanism of pyrethroid resistance observed in southern Benin is based on the modification of target in the vectors. Contrary to that observed in some African countries, such as Burkina Faso , this resistance is high in An. gambiae M and S form. The M form was the predominant population in southern Benin in general, and particularly in this study area. These results confirm those of Yadouléton et al.  showing that the resistance of malaria vectors to insecticides was growing in Benin.
In order to determine the influence of pyrethroid resistance on LLIN efficacy, the evolution of vector susceptibility in the study area was monitored. Survey results showed that phenotypic resistance varied strongly over time when compared with 2011 data . This variation has led to recommendations for the WHO village classification. Indeed, median value of the deltamethrin mortality was used for clustering of villages of high and low resistance. The median value for mortality in this study was 79%. This suggested that the mortality induced by deltamethrin has decreased. So, vector susceptibility to deltamethrin appears a dynamic phenomenon, which could be influenced either by intra- and extra-parameters, such as climatic conditions, ecological factors, or season.
Kdr mutation is responsible of pyrethroid resistance but detoxification mechanisms are also involved. Until now, the part of each mechanism does not know in the phenotypes observed in this study. Kdr results showed that there was a significant difference between the low and high resistance villages in 2012. The frequencies of this mutation are significantly lower in low resistance areas than in high resistance areas. The mutation was also found either in the M and S form. This could be explained by a high selection pressure of the kdr gene in the field populations of vectors. Therefore, the correlation between phenotypic resistance (susceptibility to deltamethrin) and genotypic resistance does not observe . The metabolic mechanisms involved in pyrethroid resistance are present in Benin [28–31]; complementary studies on these genes should be conducted to address this question.
The LLIN coverage of households in children provenance in this study (88%) and the utilization rate of LLINs by children (71%) were better. Furthermore, no significant difference was observed between the coverage and the usage of LLIN in both localities ( R + and R + + + ). Thus, both arms have been homogeneity and these factors do not affect the analysis of results.
The prevalence of malaria parasitaemia in this study population was 22%, and variations were found between clusters (5.0-57.5). It was lower than the 44.4% prevalence reported in children < five years of age from the malaria indicator survey conducted in the same region in 2010 . These prevalences were similar to those observed by Pond  among children living in rural communities distant by 150 km to cities or within the same zone of malaria endemicity. This study showed that in 14 of 20 large cities, all the children living in 75% or more of the clusters were malaria parasite-free. The decrease in the prevalence of malaria parasites may be due to the control measures recently implemented by the Benin Government through the Ministry of Public Health . The measures include a nationwide free distribution of LLINs . The decline in malaria burden attributed to the use of interventions such as LLINs was also reported in malaria-endemic countries, such as Kenya .
The prevalence of anaemia in this population of young infants was 78.9%, nearly identical to those rates reported (79%) for the region in the malaria indicator survey conducted in 2010 . The prevalence of anaemia observed in the study is not unexpected as a positive relationship with resistance. The level of haemoglobin (<11 g/dl) used as an indicator of anaemia was not significantly influenced by vector resistance to insecticide. Achidi et al. in Cameroun showed that the difference of prevalence of anaemia was not unexpected in the locality. They could potentially reflect the decline of nutritional status.
In this study, LLIN effectiveness in malaria prevention was significantly higher in the resistance area. The prevention of anaemia by the use of LLINs was also higher in areas of high resistance. According to a recent study on malaria transmission in the study area , vector density was very high in low resistance areas. These authors noted in low-resistance area a high EIR of 184.5 infected bites /man /6 months against 66.7 infected bites /man /6 months ( p <0.001) in the high resistance area. Similarly, the prevalence of malaria infection was 27.1% in low resistance area against 17.3% in high resistance area. However, no significant difference was observed between the prevalence of anemia in two areas. The high level of transmission obtained in the region should thus lead to a greater number of malaria cases. The results of a recent study  suggest that feeding on human hosts whose blood has been depleted due to severe anaemia did not significantly reduce the ability or potential transmission of malaria vectors, and indicates that mosquitoes may be able to exploit the few resources from a low level of haemoglobin rather than one that is normal in order to reproduce. For proper evaluation of the impact of vector resistance to pyrethroids on the effectiveness of LLINs, it would be desirable to have two frankly different areas of susceptibility vectors status: one where the Anopheles was resistant and another one where Anopheles was fully susceptible. In addition, the two areas must have the same ecological characteristics. Unfortunately, the sharp increase in the vectors resistance in southern Benin, has not allowed us to obtain such areas and this is what constitutes the main limitation of this study. Another limitation of this study was the cross-sectional study design. Associations presented could have been confounded by unmeasured factors and therefore causal inferences cannot be drawn. In addition, the temporal relationship between exposure variables (evolution of resistance vectors, the effectiveness of the use LLIN) and outcomes of interest (occurrence of malaria cases and other related factors) cannot be observed. Finally, because this study enrolled participants using convenience sampling and was done in a single geographically defined area, care should be taken in generalizing the results to the other populations.