Difficult decisions, such as the one to withdrawal IRS from an area, must be taken by National Malaria Control Programs when faced with demands that exceeds the availability of resources for control. Providing an alternative, proven and effective vector control intervention, such as LLINs, to replace IRS can improve, albeit not completely, the negative consequences of IRS withdrawal. The challenge is especially great in situations like Ouémé Plateau District, where the transmission season is long and the entomological inoculation rate is high . Evaluating the impact of IRS withdrawal and replacement with high coverage LLINs to inform this issue is an important objective of entomology monitoring and evaluation. The example of IRS withdrawal in Southern Benin provides a good case in point. It is costly, and difficult to implement IRS without assistance from external partners. Nonetheless, this intervention is highly effective and requires little in terms of compliance from the community. The problem with IRS is; can it be sustained independently? In the case of Ouémé there was a concerted effort to document the impact of LLINs and to reissue the population of Ouémé department, as well as the NMCP that the alternative, universal coverage with LLINs following withdrawal of IRS, would provide a similar level of protection from malaria.
However, in this case, entomological indicators of malaria transmission increased significantly after IRS withdrawal /LLIN distribution. LLINs are proven and effective, however impact also depends on the existence of a strong ‘net culture’ in the community. For example, proper use (compliance with nightly use of sleeping under nets) and care of nets is a key behavior change that must take place if LLIN interventions are to be as effective as IRS. High coverage with LLINs is also necessary. Unfortunately, in this study there was no assessment of LLIN use at the community level to evaluate how well communities adapted to the unique aspects of the LLIN.
This study suggests that in Ouémé department multiple entomologic measures of malaria transmission increased, raising the risk of transmission, at a time when a new intervention, universal coverage with LLIN, was introduced, after IRS withdrawal. The HBR increased following withdrawal of IRS and hanging of LLINs could not initially be kept during the IRS campaign despite the distribution of treated nets to every household. Vector biting density increased in three of the assessment districts after withdrawal of IRS (Adjohoun, Dangbo and Missérété). The lower HBR observed in Sèmè after withdrawal, an exception to the pattern, could be explained by the fact that the breeding sites and the An. gambiae larvae had been washed away by waters from the heavy rains in that district in 2011. Such a high rainfall was not observed in the other districts. The HBR drop also observed after IRS withdrawal in the control area, was probably due to a higher rainfall in 2010 than in 2011. This reduction of HBR obtained in the control area may be due not only to changes in rainfall but also the increase of LLIN coverage, which could contribute to the reduction of HBR and the sporozoite rate during the two periods. It can be concluded that the withdrawal of IRS and distribution of LLINs is associated with an increase in HBR.
HBR in Sèmè decreased after withdrawal, which contradicts the pattern in the other sites. However, rainfall in this district was lower after withdrawal than before. We speculate that this shift in rainfall, not seen at the other sites, may be responsible for the HBR decrease following withdrawal.
An increase in malaria transmission was observed in the districts after IRS withdrawal. In Adjohoun and Missérété, the entomological inoculation rate obtained after IRS withdrawal (9 months) was significantly higher (p<0.001) than that obtained during IRS. Such a variation in the transmission level is probably due to the high An. gambiae HBR and the large CSP index recorded after IRS withdrawal.
Compared to Adjohoun, Dangbo and Missérété, malaria transmission remained high in Sèmè during both periods. But the EIR in this area after IRS cessation was less than that recorded before withdrawal of IRS. This difference may reflect the high Culicidae nuisance problem in this area. If nuisance biting was exceptionally high, compliance with nightly use of nets may have been better that in other sites. Indeed, after IRS cessation, the use of LLINs proved successful in reducing malaria transmission in this area of high anopheline aggressiveness. This confirms the work of Kelly-Hope et al.  and Akogbéto et al., that nuisance biting improves compliance with LLIN use. Moreover, the results of research in Africa showed that vector control by widespread use of LLINs is a strategy that can reduce malaria morbidity by 50 to 60% and overall mortality by 20% [4, 25, 26].
It is also important to mention some of the limits of this study namely, the difficulty of comparing longitudinal data without a proper control area where IRS would have been maintained to control for any confounding factors such as climatic conditions or the use of mosquito nets.
The presence of Olyset nets in houses after IRS cessation did not increase the exit rate of An. gambiae and Culex spp at all. These results show that the excito-repellent effect of permethrin on An. gambiae was not remarkable, probably because of the resistance of this vector to pyrethroids. These results are reminiscent of those from N’Guessan et al. , which showed a decrease in the effectiveness of pyrethroid-treated nets, especially in Benin.