Background
Recently, the international community has prioritized national and regional malaria elimination based on the strategy of shrinking the malaria map from the margins inwards [
1]. The Solomon Islands is at the eastern edge of the distribution of malaria in the southwest Pacific, the disease is endemic throughout many parts of the country and up until the early 1990s was the leading public health problem [
2]. The Solomon Islands has now entered the malaria pre-elimination stage together with 31 other countries [
3]. To initiate this, the country's National Vector Borne Disease Control Programme (NVBDCP) has selected Temotu and Santa Isabel Provinces in which to commence malaria elimination programmes. In the remaining six provinces intensified control will be implemented.
Malaria eradication was attempted in Solomon Islands through a Malaria Eradication Programme (MEP) conducted during 1970 - 1975 [
4]. On Santa Isabel, this programme was highly successful and elimination was nearly achieved with entry into the consolidation phase (cessation of DDT - IRS) planned for 1974 [
4]. However similar success was not achieved in other parts of the country, most notably on the north coast of Guadalcanal [
5]. The MEP was abandoned in 1975 and reduced to control measures and focal DDT indoor residual spraying (IRS) to contain outbreaks. The malaria rate rose during this period peaking in 1992 [
2], in 1993 insecticide impregnated bed nets (ITN) were introduced [
6] and in early 2000, DDT was replaced by lambdacyhalothrin (ICON) for IRS, also at this time ITNs were being replaced by long-lasting insecticidal nets (LLIN). Through these intervention methods Santa Isabel Province has maintained a continual attack on the vector for the last 40 years with the introduction of ITN and LLIN resulting in a continual fall in transmission rates from 44% in 1992 to 4.5% in 2008 [Source: SI NVBDCP data: unpublished]. The current malaria situation on Santa Isabel indicates a further reduction in malaria transmission. Passive case detection at the main hospital in the capital Buala recorded 81 cases over 33 months, January 2007 to September 2009. At least 11 of these cases were reported as being imported from other provinces - Guadalcanal and Malaita. The species composition was
Plasmodium falciparum 47% and
Plasmodium vivax 53%. A mass blood survey conducted on Santa Isabel in October 2009 screened 8552 of the local population. A total of 14 were found positive by microscopy (n = 1) and PCR (n = 13) giving a parasite prevalence rate of 0.16%; of these 92.9% were due to
P. vivax and the remainder due to
P. falciparum [
7].
Several species of anophelines occur in the Solomon Islands, six members of the
Anopheles punctulatus group:
Anopheles farauti (formerly
Anopheles farauti 1),
Anopheles irenicus (formerly
Anopheles farauti 7),
Anopheles hinesorum (formerly
Anopheles farauti 2),
Anopheles punctulatus, Anopheles koliensis, and
Anopheles rennellensis [
8,
9]; as well as three members of the
Anopheles lungae complex:
Anopheles lungae, Anopheles solomonis, and
Anopheles nataliae [
10]. Of these nine species, only
An. farauti, An. punctulatus and
An. koliensis are considered vectors of malaria [
10]. With the latter two species the use of IRS and LLIN over the years appears to have eliminated
An. koliensis and
An. punctulatus is now uncommon with a patchy distribution [
11,
12]. The third species,
An. farauti, did not respond well to IRS during the MEP and subsequent control programmes due to a change in behaviour to early evening, outdoor biting [
12]. This species is now the primary vector in the Solomon Islands, being the most abundant and widespread species [
13]. It is primarily a coastal species capable of breeding in brackish water, a trait that has facilitated its spread throughout small island groups [
14].
Anopheles farauti is a member of the
An. farauti complex within the
An. punctulatus group and is morphologically indistinguishable from
An. hinesorum and
An. irrenicus, the other members of the complex found in the Solomon Islands. As
An. hinesorum and
An. irenicus are zoophilic, non-vectors species in the Solomon Islands, it is important to accurately identify these species as efforts may be wasted on controlling species of no medical importance.
On Santa Isabel, the last anopheline faunal survey was conducted between 1966 and 1971 [
11] with the species composition updated in 1978 [
13]. From these records
An. farauti s.l.,
An. punctulatus, An. koliensis, An. lungae,
An. solomonis, and
An. nataliae were recorded. However, during the past 30 years it is possible that species composition and distribution may have been altered by continual pressure from the vector control programmes as well as biotic factors such as competition and dispersal of zoophilic species. Moreover, as these surveys predate the use of molecular techniques for species identification it is possible that other species may occur on the island. For example by using allozymes,
An. hinesorum was identified from two locations around Buala in 1988 [
15]. With the molecular based techniques currently available that will reliably identify the members of the
An. punctulatus group the presence and distribution of these species on Santa Isabel can now be resolved; however no such techniques are as yet available for identifying the members of the
An. lungae complex.
The strategies planned for malaria elimination in Santa Isabel Province are improved and more widely available diagnosis, greater accessibility to more effective treatment drugs, and vector control with LLIN and IRS aiming at reducing human vector contact, vector longevity, and vector density. To support the vector control measures information on the speciation, distribution, ecology, behaviour, and biology of the vector is important to ascertain what type of elimination measures are most appropriate and for which areas. The results of these surveys will also provide base line data for future monitoring and evaluation of the elimination programme. As this information is currently lacking for Santa Isabel entomological surveys were conducted in October 2009. This paper reports on the findings of these surveys
Methods
Study site
This study was conducted in Santa Isabel Province (8°14'21.66″S latitude and 159°33'27.08″ E longitude). The province is made up of the main Santa Isabel Island (200 km long by 25 km wide) and several closely associated smaller islands. A mountain range, up to 1500 m above sea level (asl), runs down the center of the island and there is a narrow coastal shelf (0.1 - 3 km wide and <20 m asl) that runs round the island; the majority of the population live in villages that lie along this coastal shelf. The climate is continuous hot/wet. There is no rainfall data specific to the province but weather stations in surrounding provinces indicate that rain occurs all year round and at a rate of about 2,500-3,000 mm per annum. The mean temperature for the region is 26-27°C on the coast and inland lowland regions. The total population is 26,500, most of whom live on the eastern end of the island around Buala the provincial capital. There is no road network and all movement is by boat.
Species composition and distribution
Utilizing marine transport a larval survey was conducted covering most of the island. As the majority of the villages were located on the coast, the survey concentrated on this area; however inland villages were also surveyed. At each location, using local knowledge from the village people, all ground pools were checked around the villages. Using standard 250 ml dippers, larval samples were taken from all sites positive for anopheline larvae. A description of the site was made and its location geo-referenced on 1:50,000 scale maps. The anopheline larvae collected from each site were preserved in 70% ethanol in appropriately labeled vials.
Selection of indicator villages for adult collections
Two indicator villages - Kolosori and Popoheo - were selected for studying the biology and behaviour of the adult anopheline fauna. Kolosori (8°07'13.21″ S, 159°31'49.96″E) is located on the coastal plain about 3.0 km inland and 20 m asl and Popoheo (8°05'34.18″ S, 159°31'24.10″E) on the same plain but directly on the coast (<5 m asl) and surrounded by brackish water swamps. Both villages were chosen in the vicinity of Buala for reliable access. Three houses in each of these villages were randomly selected for the indoor human landing catches. In Kolosori a host preference study was carried out comparing the attractiveness of paired human and pig baits.
Human landing catches (HLC)
Local villagers from Buala were hired and trained to perform the human landing catches. For each indicator village, 12 collectors were hired, with six working from 6 pm to midnight and the other six working from midnight to 6 am. Collections were made for 10 nights in each indicator village.
Three collectors worked indoors and three outdoors. Each collector collected for 40 min each hour; during this period the collectors, using a torch and aspirator, caught all anophelines coming to feed on their exposed legs and feet. Specimens collected were held in separate cups labelled for each hour, either outdoors or indoors. The collectors were alternated with regards to indoor/outdoor and collection time either the first or second parts of the night. The cups were covered with damp cotton wool and held until the following morning when the mosquitoes were killed using chloroform, morphologically identified [
10] and the hourly collection rates recorded. All adult mosquitoes collected were preserved, desiccated on silica gel, in microfuge tubes each labelled with the village, date, time of feeding, method of collection and whether collected indoor or outdoor.
Longevity of vector populations
Parity was determined based on the condition of the tracheolar skeins of the ovaries [
16] this was done on all anophelines collected each hour from the night landing collections. From this the proportion parous (P) was used to determined the survival through one day (
p) as
x
√P where
x is the length of the gonotrophic cycle. The proportion of the vector population living long enough to transmit malaria parasites was determined by
p
n
where n was 9 days for
P. vivax and 12 days for
P. falciparum [
17].
Human host behavior
A census of the people outdoors each hour from 6 pm - 6 am was conducted in each of the indicator villages on each night that the human landing catches were performed. Based on the total number people in the village, the number of people indoors each hour was then determined and related to the biting times of the anophelines.
Host seeking in relation to host indoor/outdoor movements
The use of indoor and outdoor collectors creates an artificial situation where for example mosquitoes collected outdoors, late in the night when no village people are outdoors, are counted as outdoor biting. It is more realistic when estimating the degree of indoor (or outdoor) biting to account for the spatial interaction between the vector and the host, resulting from the movement of vectors and hosts in and out of houses. This is particularly so where the vector control tools (IRS and LLIN) being implemented only target indoor biting mosquitoes. The degree of indoor or outdoor biting will be a factor of where the people are at the time of vector host seeking. Thus the proportion of indoor biting mosquitoes (π
i) was calculated by multiplying the number of
Anopheles biting indoors with the number of people indoors for each hour of the night and similarly the outdoor biting component (π
o) was multiplied by the number of people outdoors for each hour of the night. The mathematical formulae for estimating π
i have been detailed previously [
18].
Molecular identification of mosquitoes
All specimens collected were initially identified using morphological keys [
10]. As many of the
Anopheles species in the Solomon Islands belong to isomorphic species complexes the morphological identification of the specimens collected was verified using polymerase chain reaction - restriction fragment length polymorphisms (PCR-RFLP) of the ITS2 region of the ribosomal DNA [
19]. Up to 10 individual larvae from each larval collection site, and all of the adults from the human landing catches and the animal baited traps were processed. DNA extraction, amplification, restriction digest, fragment separation, and visualization are as previously described [
20]. With any anomalies or unknown RFLP then the ITS2 was sequenced for that specimen and compare to the existing sequences in the GenBank [
20].
Vector incrimination
The head and the thorax of all adult mosquito specimens collected biting humans and from bait net traps were processed using specific monoclonal antibodies and an enzyme-linked immunosorbent assay (ELISA). The methods used were those supplied with the monoclonal reagents (Dr Robert Wirtz, Centers for Disease Control and Prevention, MS F42, Atlanta, GA 30341-3717, USA). Specimens were considered positive if the absorbance value was twice that of the average negative control value and all positives were rerun for confirmation.
Animal host preference
Two animal baited traps were set up, one containing a human, the other two small pigs. To protect the human bait the person slept inside a normal untreated bed net covered by the animal bait-net. With both the human and pig traps, the skirt of the nets were raised 10 cm off the ground at 6 pm each evening to allow entry of blood seeking mosquitoes during the night and lowered at 6 am each morning. The traps were inspected at 7 am each morning and all trapped mosquitoes were collected, counted and scored for unfed or blood fed.
Ethics
Ethical approval for the study was obtained from the University of Queensland Medical Research Ethics Committee (2010000412).
Discussion
A major finding of this work is the absence of
An. punctulatus and
An. koliensis from Santa Isabel. These two primary vectors of malaria were recorded on the island in 1970, at the time
An. punctulatus was uncommon, but
An. koliensis was common and widespread [
8]. Their disappearance is most likely due to years of IRS, first with the MEP and then the on-going control programme. This occurred in New Guinea and in other parts of the Solomon Islands when IRS was introduced, though with the cessation of the eradication programmes or a reduction in IRS coverage they rebounded [
11,
12,
23,
24].
A further factor which may have helped with their elimination is competition from
An. hinesorum; which was found to be dominant anopheline species on Santa Isabel.
Anopheles hinesorum occupied breeding sites commonly used by
An. koliensis, particularly drains and semi-permanent ground pools. It was also commonly found in transient ground pools (Figure
3). These sites are normally used exclusively by
An. punctulatus [
14] to the point where when
An. farauti and
An. koliensis larvae were seeded into these types of sites they failed to develop [
25].
Anopheles hinesorum populations in the Solomon Islands are zoophilic and are not involved in malaria transmission [
26,
27]. The disappearance of
An. punctulatus and
An. koliensis and their replacement by a zoophilic, non-vector species is likely responsible for some of the decline in transmission rates, particularly in inland villages.
Anopheles farauti, the third major malaria vector in the Solomon Islands, was common on Santa Isabel, it was found predominantly on the coast where the majority of the human population is located. The larvae of
An. farauti will develop in brackish water and as found here in Popoheo and recently in coastal villages in Temotu Province large brackish coastal swamps are the most productive sites for this species; these site can be responsible for high human biting densities [
21].
In this study,
An. farauti was incriminated as a vector for vivax malaria in Popoheo, and is most likely the main vector on Santa Isabel. There have been problems in the past with the control of this species. Of concern is its shift to early night outdoor feeding thought to be due to an excito-repellent response to the IRS. This change in biting behaviour allowed it to avoid the insecticide and maintain transmission outdoors limiting the effectiveness of past eradication and control programmes [
5,
12,
28]. In Popoheo
An. farauti was found to be an early night feeder with 52.3% of host seeking occurring between 6 - 8 pm, however it was still quite endophagic and readily enter houses to feed even during the early hours of the night when considerable numbers of villagers were outdoors. This is anomalous with
An. farauti populations in other parts of the Solomon Islands where similar intervention measures resulted in outdoor/indoor feeding ratios of 1:0.25 [
12,
15]. Whether on Santa Isabel the withdrawal of IRS, with only occasional focal spraying, has removed the insecticide pressure allowing for a return to indoor biting is not known. Adjusting outdoor and indoor biting for the movement of people from outdoors to indoors during the night, it is estimated that 54.6% of the biting of
An. farauti occurs indoors each night and that this proportion of the vector population could be controlled if good coverage of IRS and LLIN was achieved in the implementation of the elimination programme.
In the inland village of Kolosori,
An. solomonis was recorded as the dominant human biting anopheline. This is the first time this species has been found biting humans in appreciable numbers, prior to this there was very little evidence for considering this species as a possible malaria vector [
10].
Anopheles solomonis had a pronounced early night feeding pattern and it was extremely exophagic. This type of behavior would make this species difficult to control with IRS or LLIN. The results of the host preference study indicated that
An. solomonis would readily feed on pigs (4.6/night) with no specimens collected off the paired human bait, though at the same time the night biting catches recorded an overall biting rate of 4.75 b/p/n. A mass blood survey conducted in Kolosori village at the same time as the entomological survey found no positive malaria cases and all (285) human biting
An. solomonis assayed for circumsporozoite antigen were negative.
Kolosori was the only place in the Solomon Islands to date where
An. hinesorum has been found biting humans; this was in very low numbers (2.96% [9/304]), particularly in light of the number of larval sites positive for this species (13 sites) in the immediate vicinity of Kolosori village. This species is inconsequential with regards to malaria transmission. These finding support previous work on this species in the Solomon Islands where it has been identified as a zoophilic species [
26,
27].
The malaria vector
An. farauti was also found biting humans in inland Kolosori but in small numbers (3.29% [10/304]). The flight range for this species is up to 3 kilometers [
29], and in the absence of
An. farauti larvae in the Kolosori area it is most likely that these few specimens have come in off the coast. From a malaria transmission point of view, these low numbers are unlikely to contribute to malaria transmission, though the possible continued presence of this species in these inland villages should be monitored. The proportion parous was low for both
An. farauti and
An. solomonis, being 0.41, and 0.33 respectively. These values are indicative of short-lived populations with only a very few specimens living long enough to transmit the parasite. For
An. farauti in other parts of its range in the Solomon Islands and Papua New Guinea considerably higher survival rates, ranging from 0.55 to 0.80, have been recorded [
4,
30].
Unlike the provinces of Guadalcanal and Malaita, where malaria is more intense, the disease does not seem to be well entrenched on Santa Isabel. Since the MEP of the early 1970s, continued use of focal IRS, and the introduction of insecticide treated nets in 1993, has seen a decline in the malaria cases with an overall parasite rate of 0.26% being recorded in 2009. The mass blood survey conducted in October 2009 estimated a similarly low overall parasite rate of 0.16% [
7]. The reason for the decline in malaria rates is most likely to be due to the elimination of
An. koliensis and
An. punctulatus from Santa Isabel leaving
An. farauti as the only recognized vector. This species can be an efficient vector of malaria and in villages on Guadalcanal it has been responsible for maintaining parasite rates of 30-46% [
30]. However on Santa Isabel, despite reasonable biting densities at Popoheo, the sporozoite rate was 0.00129 and only two human cases were found in the village during the mass blood survey of 2009 [
7]. It is possible that there is sufficient indoor host-seeking to result in some degree of
An. farauti exposure to insecticide treated nets or IRS which may be responsible for the low survival rate of this species on Santa Isabel and thus its poor vectoring ability. It was found on Guadalcanal, that the introduction of insecticide-treated nets resulted in a parous rate in
An. farauti of 0.39 as compared to 0.55 in a no-net site and 0.57 in a DDT IRS site [
30].
With An. farauti restricted to the coast and the disappearance of An. koliensis and An. punctulatus, the inland villages (>3 kilometers from the coast) of Santa Isabel should be malaria free unless An. solomonis plays some role in transmission. Even if this species was capable of transmitting the parasites, its low biting densities, a degree of zoophilia, and very low survival rates would make it a particularly inefficient vector.
In the Solomon Islands, factors that have been shown to play a lead role in the decline of malaria rates are IRS, LLIN, and education [
2]. On Santa Isabel all three of these factors are currently implemented. Additionally the communities on Santa Isabel have a national reputation for a high level of active participation and support for self-betterment. Such community attitudes promote high levels of acceptance for activities like malaria awareness and education plus support and good co-operation for the implementation of IRS and LLIN intervention measures.
On Santa Isabel, malaria elimination is being sought through LLIN distribution to all households in all villages throughout the province in addition to complete and thorough IRS coverage in all houses in all villages along the coast targeting the distribution of the main vector An. farauti. This is a realistic approach as IRS is very labour intensive and costly and just focusing on the main vector allows for a concentration of resources where most needed. Additional to vector control there will be better access to diagnosis and treatment with the recent introduction of a new treatment drug Coartem® (artemether + lumefantrine) to replace chloroquine which has become increasingly ineffectual due to P. falciparum and P. vivax resistance. The effect of these measures on the already declining parasite rate should see the achievement of malaria elimination from Santa Isabel though the impact of the IRS and LLIN on An. farauti will have to be monitored as will the possible role An. solomonis might play in transmission in inland villages.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
Conceived and designed the experiments: HB, RDC, AB. Performed the experiments: HB, RDC, TLR, SC, DOM, CI, AA, WH. Performed the molecular analysis: NB, DOM. Analyzed the data and wrote the manuscript: HB, RDC, TLR. Reviewed the manuscript: AB, DOM, TLR, CCC. All authors have read and approved the final manuscript.