Previously published laboratory based evidence showed that colonized An. gambiae s.s. is susceptible to IVM at concentrations relevant to human pharmacokinetics after a typical MDA , and that colonized An. gambiae s.s. fed on IVM-treated humans one day post-treatment had reduced survivorship . The current study now demonstrates that routine MDA of IVM to people significantly reduces the survivorship of wild An. gambiae s.s. for up to six days post MDA. This six day lethal effect is longer than the two days observed from Kobylinski et al. , and this effect occurs despite incomplete MDA coverage in treated villages. Three field replicates were performed over space and time in different villages to make this study a rigorous assessment of the effects of IVM MDA on An. gambiae s.l.
There were no significant differences in An. arabiensis treatment by phase survivorship (F-value = 0.66, P = 0.5332) but this was almost certainly due to sampling, in that adequate numbers (n = 153) for survivorship analysis were only captured in the third replicate MDA. The third replicate MDA was performed in October 2009, toward the end of the rainy season when An. arabiensis is more prevalent . There was a 38% reduction in mosquito survivorship from phase 2 treatment collections compared to phase 2 control collections (Figure 4). The fact that treatment alone was significant (F-value = 7.01, P = 0.0191), means that the overall survivorship of An. arabiensis was lower in the treated village compared to the control village (Figure 4). Furthermore, only 75% (24/32) of An. arabiensis blood meals were from humans (Figure 5), which reduces the probability that mosquitoes held for survivorship analysis may have ingested an IVM-containing blood meal. Fritz et al  reported that colonized An. gambiae s.s. and colonized An. arabiensis have almost identical susceptibility to IVM. When these data are considered together, it is reasonable to assume that upon further replication wild An. arabiensis will be shown to be as susceptible to IVM MDA as wild An. gambiae s.s.
Based on clinical records, 84.2% (203/241) of people in Damboucoye and 82.1% (311/379) of people in Ndebou were treated with IVM during these two MDAs. Pregnant women and children under 90 cm did not receive the drug, following the manufacturer guidelines. Mosquitoes that were held for survivorship analysis for five days had completely digested their blood meals, which made it impossible to detect IVM from individual mosquitoes. Yet it is impressive that mosquito survival was still significantly reduced despite not knowing whether any one mosquito fed on a treated person. Incomplete coverage may actually be beneficial to the overall concept of repeated IVM MDAs for malaria control in that it may provide a refugia of untreated human hosts for mosquitoes to feed on which could reduce the likelihood of IVM resistance development in the mosquito population.
Of people accounted, 78.2% (903/1,155) utilized ITNs across the four villages surveyed in 2009. Even with high ITN coverage, human blood fed An. gambiae s.s. and An. arabiensis were frequently collected from the inside of huts, demonstrating that ITNs have limitations in preventing Anopheles from feeding on people in huts. Exophagic and exophilic malaria vectors also comprise an important part of the malaria transmission cycle in this study area; Anopheles funestus group mosquitoes are almost twice as likely to blood feed outdoors than indoors , and Anopheles nili tend to be exophilic or immediately exit huts after biting [20, 27]. ITNs may reduce malaria transmission by exophagic vectors [28–30], but their primary efficacy is against endophagic vectors. It has also been shown that ITNs may shift vector host seeking times to earlier in the evening when people will not be sleeping under an ITN . IRS will only affect the survivorship of endophilic vectors that contact the sprayed surfaces, and it is believed that the exophilic portion of the An. gambiae s.l. population led to the failure of IRS to eliminate malaria transmission during the Garki project . Furthermore, a number of malaria vectors will naturally feed at dusk and dawn, when humans are less likely to be indoors and protected by an ITN or by IRS. Ivermectin MDA may be one of the few methods that can directly target these exophagic, exophilic, and crepuscular-feeding malaria vectors, and should integrate well with the employment of existing in-home control methods like ITNs and IRS.
Ivermectin has a different mode of action from the insecticide classes currently used for ITNs and IRS (i.e., carbamates, pyrethroids, and organochlorines) , in that it agonizes the glutamate-gated chloride anion channels found in invertebrate postsynaptic neurons and neuromuscular junctions [33, 34]. This action hyperpolarizes the neurons and muscle fibers, leading to flaccid paralysis and insect death [35–37]. Once or twice yearly IVM MDA has been occurring in this region for over fifteen years , so the fact that a reduction in survivorship of An. gambiae s.s. was still detectable is a promising sign that resistance by Anopheles spp. may be slow to develop against this drug. Furthermore, the novel mode of action of ivermectin compared the currently used insecticides for malaria control should potentially minimize issues of cross-resistance where IVM MDA may be used in combination with IRS and ITNs.
Anopheles gambiae s.s. often requires two blood meals to complete its initial gonotrophic cycle [38, 39] and thereafter will often take multiple blood meals per gonotrophic cycle [40–42] and feeds almost exclusively on humans (Figure 5) . These blood feeding characteristics, coupled to the fact that the extrinsic incubation period for Plasmodium spp. is 9-14 days, means that most malaria transmission by An. gambiae s.s. will occur only after Plasmodium parasite-harbouring mosquitoes have taken multiple non-sporozoite transmitting blood meals from humans [21, 43]. If human population clusters were simultaneously treated with IVM MDA, then most adult An. gambiae s.s. in the MDA area would imbibe a concentration of IVM that would reduce their survivorship. The predicted effect has the potential to temporarily shift the An. gambiae s.s. population age structure, which would reduce the reservoir of adult sporozoite-transmitting An. gambiae s.s. in the MDA area. The low sporozoite rate in the resulting mosquito population would temporarily reduce the basic reproductive number (R0) of malaria below the base number for approximately eleven days post IVM MDA. Current IVM MDAs for onchocerciasis control in Africa are performed only once or twice per year, and do not always coincide with local malaria transmission seasons. Such current applications would not be expected to lower malaria transmission long enough to see any noticeable reductions of parasite prevalence, intensity or disease in people. Indeed, malaria is hyperendemic in APOC-control areas of south-eastern Senegal despite IVM MDAs for more than 15 years. However, if IVM MDA is administered repeatedly, R0 can be reduced for an extended period of time.
This model, like all models, makes assumptions that may not be realistic in nature, such as homogeneous mosquito biting, no spatial structure, and the lack of density-dependent effects. However, the model results are conservative, since it only incorporates the direct IVM-related mortality effect observed in the field. Kobylinski et al  demonstrated in the lab that multiple sub-lethal IVM containing blood meals compounds mosquito mortality, but the model assumes complete recovery of surviving mosquitoes three days after ingesting an IVM-containing blood meal. Kobylinski et al  also demonstrated that the re-blood feeding frequency and blood digestion of An. gambiae s.s. are delayed after imbibing relevant IVM concentrations. Finally, Fritz et al  and personal observations have witnessed mosquito knockdown effects immediately after drug ingestion. All of these negative sub-lethal effects would probably lead to a further reduction in An. gambiae s.s. survivorship in the field, due to desiccation, predation, or insufficient nutrition, beyond that of outright mortality induced by IVM. Thus, it may be that the age-structured model is an underestimate of the true effect of IVM MDA on malaria transmission.
The age-structured model predicts that strategically administered, repeated IVM MDAs would achieve sustained reductions in malaria transmission. Such repeated MDAs may only be logistically feasible for areas with seasonal malaria transmission, such as the Sahel, or in areas experiencing malaria epidemics. An expanded IVM MDA regimen fits well with the idea of combating polyparasitism in some of these same communities [44, 45]. Malaria and soil-transmitted helminths (STHs - Ascaris lumbricoides, Trichuris trichiura, Strongyloides stercoralis, and hookworms) are co-endemic across much of sub-Saharan Africa [46–50]. The annual/biannual IVM MDAs for onchocerciasis control are not generally sufficient for controlling STHs because of relatively rapid re-infections due to their transmission dynamics [51–55]. Repeated IVM MDAs are likely to result in reductions of the prevalence and intensities of STHs in individuals receiving the drug [52–59], and would increase the personal incentive to participate in such MDAs. The combination of malaria and soil-transmitted helminth infections can exacerbate anaemia, resulting in worsened child development and more adverse pregnancy outcomes than these diseases cause on their own [60–64]. Therefore, the combined effects of IVM MDA on multiple parasites could potentially reduce anaemia in the human population which would lead to an overall improvement in human health beyond what would be expected from malaria control alone.
Future field work will need to be conducted to determine if repeated IVM MDAs can quantifiably reduce malaria transmission and if the model created here is an over or underestimate of what would occur in the field. Important to this future work is determining whether a logistically feasible IVM MDA interval can be devised that would reduce malaria transmission below a critical threshold and not foster IVM resistance in either Anopheles vectors or human nematode parasites.