Background
The basic reproductive number (R
0) [
1] is determined by the intensity of malaria transmission which depends largely on the parameters comprising vectorial capacity [
2,
3] (the human biting density, proportion of blood-meals on humans and the mosquito life expectancy). The vector life expectancy, in turn, is a function of its survivorship per feeding cycle and the length of the feeding or gonotrophic cycle. The effectiveness of vector control depends on where and when a vector seeks human blood meals (which is determined, in part, by the location of humans). These parameters vary by species, both geographically and temporally, and will determine the effectiveness of vector control strategies implemented across different seasons and locations.
The Solomon Islands is currently undertaking country-wide intensified malaria control with the goal of malaria elimination in targeted provinces. Malaria transmission is predominantly by
Anopheles farauti. The vector control strategies are those recommended by the World Health Organization’s Malaria Policy Advisory Committee (WHO MPAC)—universal distribution of long-lasting insecticidal nets (LLINs) and indoor residual spraying (IRS) in limited areas [
4]. Exposure of vectors to the insecticides occurs when mosquitoes enter houses late at night while seeking a blood meal (LLINs) or when resting after blood feeding (IRS) [
5,
6]. Although
An. farauti varies in its degree of anthropophagy across Melanesia, it is highly anthropophagic in the Solomon Islands [
7]. In the Solomon Islands,
An. farauti displays behavioural resistance to insecticides by feeding mostly outdoors and early in the evening [
8].
Anopheles farauti first shifted its behaviour to feed early in the evening when people were outdoors in response to the DDT spray campaigns in the 1970s, thus avoiding the insecticide [
9,
10]. This behavioural shift was one reason that the original Malaria Eradication Programme (MEP) of the early 1970s failed [
11] and, malaria cases surged throughout the Solomon Islands until insecticide treated nets (ITNs) and LLINs were introduced in 1992–1993, and 2005, respectively [
12]. This insecticide avoidance behaviour appears to be maintained by the widespread use of LLINs, as recent surveys show that this early outdoor biting behaviour still persists in at least three other
An. farauti populations in the Solomon Islands [
13‐
15].
Despite the challenge of behavioural resistance in
An. farauti, there have been significant reductions in malaria achieved in the Solomon Islands in the past 20 years with ITNs, IRS and improved anti-malarials. However, malaria elimination remains, perhaps, an insurmountable challenge with these available intervention tools. New vector control interventions are needed to complement the indoor killing of LLINs and IRS by attacking outdoor feeding or other behavioural vulnerabilities of
An. farauti [
16,
17]. Rational development of such tools requires detailed knowledge about the biology and behaviours of vectors. The isolated island populations of
An. farauti display variability in their night biting profile, blood feeding patterns and the degree of endophily, likely the result of restricted gene flow among island populations [
18]. In this paper, a number of key vector parameters were measured for
An. farauti, in Central Province, Solomon Islands to determine potential behavioural vulnerabilities for vector control. These parameters were the daily and seasonal biting behaviour, the time and location (indoors or outdoors) of blood feeding and the length of the feeding cycle.
Discussion
The effectiveness of vector control is a function of both mosquito and human behaviours. For LLINs and IRS, the degree to which the vector feeds or rests indoors (i.e., how endophagic or endophilic) as well as the frequency at which the vector blood feeds will largely determine the proportion that survive for the duration of the extrinsic incubation period. Indoor feeding and resting are determined, in large part by the location of humans (indoors or outdoors) when mosquitoes are seeking blood meals (e.g., mosquitoes seeking human blood meals earlier in the evening are more likely to feed on humans outdoors when few people are inside houses). The duration of peak mosquito density is important for the selection and timing of the application of insecticides used in IRS (as different insecticides and formulations vary in their effective half-life).
Most populations of
An. farauti in the Solomon Islands bite outdoors and early in the evening. Previously reported π
i values (the proportion of feeds on humans taking place indoors) were 0.314 for Guadalcanal Province in 2007-08 [
15]
1 to 0.368–0.570 for Temotu Province in 2008–2010 [
13] with the highest value recorded in Isabel Province in 2009 (0.546) [
14]. The lowest proportion of bites on humans indoors for
An. farauti was found in this study in Haleta village on Ngella Sule, Central Province, with only 13 % of human feeds indoors. This island was designated as a “problem area” during the original malaria eradication programme [
10], which is understandable as the early outdoor feeding of
An. farauti found in this study would minimize exposure to the insecticides used in IRS and ITNs and limit the effectiveness of the interventions.
The terms gonotrophic and feeding cycle are often used interchangeably despite the fact that they are, in fact, describing slightly different time intervals (i.e., the period between successive oviposition and blood feeding events, respectively). Mark-release-recapture experiments using HLC estimated the feeding cycle length whereas the gonotrophic cycle length was estimated by measuring the duration between blood feeding and oviposition of mosquitoes held under field laboratory conditions. Feeding cycle length estimates from mark-release-recapture, for all anopheline species range from 2 to 4 days [
33,
34]. The feeding cycle length for
An. farauti in Central Province is one of the shortest recorded at 2.1 days, but is comparable with previous estimates for this species from Guadalcanal Province, Solomon Islands [
35] and Madang Province, Papua New Guinea [
29,
30,
36] which ranged between 2 and 3 days. The feeding cycles among malaria vectors in different villages in Madang, Papua New Guinea were 2.7–3.7 days for
Anopheles
punctulatus, 2.4–3.2 days for
Anopheles koliensis and 2.1–3.0 days for
An. farauti [
30]. The local environment was found to exert a greater influence on the duration of the feeding cycle than the species of mosquito, with permanent pool breeders having a shorter cycle then temporary pool breeders. If extensible to the Solomon Islands, the environmental conditions in the coastal villages where
An. farauti is found would have been predicted to have a short gonotrophic cycle, since the vector is laying its eggs in a permanent breeding sites (coastal lagoons and swamps) located in close proximity to villages and thus the human host.
The estimated length of the gonotrophic cycle (2.6 days) was longer than the estimate of the feeding cycle (2.1 days) calculated from the mark-release-recapture experiment. It is possible that the laboratory conditions (e.g., sugar deprivation, limited space, temperature, etc.) in which the gonotrophic cycle was estimated from egg development were sufficiently different from the field conditions in which the feeding cycle was measured to explain the difference between the estimates of the gonotrophic and feeding cycles. A similar study for
Anopheles albitarsis in Brazil [
37], also found a longer gonotrophic cycle (calculated from oviposition observations) than the feeding cycle (from mark-release-recapture experiments).
The
An. farauti population in this area exhibited a single peak biting season between October and January. In Haleta the parity data followed a seasonal trend with higher parity rates occurring during peak adult densities and declining from February with lowest rates in August and November 2012 when
An. farauti densities would begin to increase (Fig.
3) with the emergence of nulliparous mosquitoes into the adult population. This should be considered when planning vector control, with the bulk of activities completed before commencement of the peak biting season. A very similar temporal pattern and similar genetic population of
An. farauti [
18] was found in Guadalcanal [
15]. A supporting study of the larval populations in Guadalcanal [
38] demonstrated that larval presence and density also varied seasonally and was primarily driven by rainfall.
Historical estimates of the sporozoite rates and EIR for
An. farauti are not available for Central Province, but are available for Guadalcanal Province (the nearest province). During the early 1990s and in the absence of vector control, EIR values as high as 1022 ib/p/y were recorded in Guadalcanal [
39]. The intensified vector control programme implemented by the Ministry of Health and Medical Services over the last decade has had a substantial impact on transmission as evidenced by the greatly diminished and now relatively low EIRs estimated here in 2012 (2.5 ib/p/y) and 2013 (35.7 ib/p/y).
Despite the early and outdoor biting habits of
An. farauti, the frequency of blood feeding by this species offers an explanation for the substantial malaria control that has been achieved by LLINs and IRS in the Solomon Islands. With each successive feeding cycle there is a multiplicative effect that increases the proportion of the total vector population exposed to insecticides. In the Solomon Islands where the annual mean temperature is ≈26 °C, the length of the extrinsic incubation period (EIP) is estimated to be 12 and 9 days for
Plasmodium falciparum and
Plasmodium vivax, respectively [
40]. With an estimated feeding cycle of two days,
An. farauti, will have 6 and 5 opportunities to enter a house before completion of the
P. falciparum and
P. vivax EIP, respectively. Although only 13 % (π
i) of
An. farauti will be potentially exposed to insecticides by biting late and indoors during each feeding cycle, this will cumulate in significant mortality across the multiple feeding cycles required to complete the EIP. Assuming that LLINs have the potential to kill 80 % of those mosquitoes that enter and attempt to feed on sleeping humans, this could translate into 47 and 41 % population-level mortality before
An. farauti would be infectious to humans with
P. falciparum and
P. vivax, respectively.
2 This emphasizes the fact that although the population of
An. farauti is primarily exophagic, indoor vector control tools still provide significant control [
41]. This is an important consideration, as evidence has been emerging from other anopheline populations that the proportion of feeding indoors is diminishing, such as for
An.
funestus in Tanzania [
42], Benin [
43] and Senegal [
44] as well as
An.
gambiae s.s. in Equatorial Guinea [
45].
Authors’ contributions
TLR supervised the overall field studies and drafted the manuscript. TLR, NFL, FHC and TRB contributed to the experimental designs. All authors participated in the field experiments. WKC, RDC and NWB conducted the molecular analyses. All authors read and approved the final manuscript.