Background
Malaria is endemic in many regions of East Africa where climate and environment together present conditions suitable for malaria vectors and parasites [
1]. The main vector species in western Kenya are
Anopheles gambiae Giles
sensu stricto An. arabiensis Patton and
An. funestus Giles.
Anopheles gambiae and
An. arabiensis are commonly found in clear sunlit pools of water, man-made shallow water bodies, in polluted water and along the shores of large water bodies such as Lake Victoria [
2‐
8].
Anopheles funestus prefers rather permanent water bodies (8). In the last decade an increasing number of cases of malaria in formerly malaria-free areas and highland areas have become common [
9‐
11]. Several hypotheses have been proposed to explain the increased malaria transmission in the highlands, including land-use changes, global climate changes, increased drug resistance, cessation of malaria control activities, and demographic changes [
11‐
13]. Cox [
14] estimated that 34 million individuals were at risk of malaria in the East African highlands. In these highlands, transmission is probably much more focal in its distribution than in many lowland areas, as breeding sites are more common in the valley floor than on the steep valley slopes [
15]. In addition, studies report that human activities in these highlands have subsequently created potential mosquito breeding habitats [
7,
11,
16].
In the Ugandan highlands, the elimination of papyrus swamps created a habitat for
An. gambiae and
An. funestus, leading to increased malaria transmission [
11]. In the highlands of western Kenya,
An. gambiae was found only in cultivated farmland habitats but not in original forest and swamp habitats [
16]. These differences in larval distribution were attributed to the fact that farmland habitats received more sunlight, and hence water temperatures were conducive for
An. gambiae breeding. In Ethiopia, changes in land use and climate expose the highland areas to unexpected malaria epidemics, presumably due to expansion of environmental conditions suitable for malaria transmission [
17].
Research on malaria in the highlands has mainly focused on the development of early-warning systems to identify when epidemics are expected [
18‐
20] and on the effects of changes in climatic variables [
13,
21,
22]. The core idea behind these systems is that when parameters indicate a malaria epidemic is likely, resources can be channeled to prevent or contain the epidemic [
18]. However, in sub-Saharan Africa, malaria epidemics arise suddenly in mostly remote, disadvantaged settings without effective alert systems [
23]. In resource-limited countries such as those of highland East Africa, an all-or nothing approach to interventions such as insecticide spraying or bed net distribution, often results in complete coverage for some areas and no coverage for others when funds run out [
18]. Thus, regular vector control activities targeted at the malaria risk areas are more cost effective than emergency interventions that often face delays in mobilization [
24]. In addition, because full coverage of control measures is hardly achieved, integration of larval source management (LSM) into Integrated Vector Management (IVM) program will be advantageous to the fight against malaria.
In western Kenya highlands, for instance, since the implementation of the roll back malaria initiative [
25], malaria control has been based on insecticide treated nets (ITNs), indoor residual spraying (IRS) and the use of anti-malarial drugs for the treatment of malaria parasites. Following the adoption of RBM, there are indications that malaria morbidity and mortality is on a decline as a result of scaled up use of ITNs [
26,
27] and increased availability of antimalarial medicines [
28]. However, with increased use of interventions targeting indoor resting mosquitoes, the vectors are bound to develop evading mechanisms or even change their biting behavior. Exophily of the commonly known endophilic species has recently been reported [
29,
30], in addition to development of resistance in the malaria vector and parasites. There is need development and integration of complementary tools to target outdoor vectors [
31,
32].
Microbial larvicides have been proven efficient in the control of anopheline mosquito larvae and the reduction in adult mosquito densities [
33‐
36]. However, access to microbial larvicides is still a challenge for developing countries, thus calling for development of alternative larval control strategies that can utilize locally available resources. In the current study, an integrated larval source management comprising of habitat manipulation, source reduction in comparison to the application of microbial larvicides and the use predatory fish were used. The hypothesis being habitat manipulation and source reduction are as effective as the application of
Bti and the use of predatory fish for mosquito larval control.
Discussion
The results obtained indicate that the LSM intervention strategies used were effective in reducing the development of late instars larvae as few to none were recorded especially from intervention habitats. Non-intervention habitats on the other hand recorded both Anopheles and Culex larvae throughout the study in Lunyerere with the exception of Fort Ternan where no larvae was recorded during some of the months. An integrated LSM approach using environmental management through application of Bti, source reduction, habitat manipulation and the use of predatory fish showed great potential in preventing development of mosquito larvae in man-made habitats within the study area. Strategies such as the use of arrow root plants and drainage can be applied by the local communities for the control of mosquitoes in the respective study areas as they compare well with the use of predatory fish and the application of Bti. Culicine mosquitoes, mainly nuisance biters were also reduced in the intervention habitats.
Anopheline mosquito species breed in a variety of habitats; however, those created by human activities may be of particular importance for malaria transmission [
6,
7,
39]. The bio-larvicide
(Bti) although efficient in controlling mosquito larvae, it is not locally available and hence not accessible by the local community members. Therefore
Bti was used alongside environmental management through habitat manipulation by use of shade provided by arrow roots and source reduction by drainage. The results show that in Lunyerere both early and late instar anopheline larvae were greatly reduced or absent in habitats shaded by the leaves of arrow roots plants and those that were drained. The abundance of
Anopheles early instars within habitats treated with
Bti and non-intervention was not different. Indicating that ovipositing females still lay their eggs in habitats containing
Bti and as a result of feeding on
Bti contained in water, the larvae then dies before reaching late instar increasing
Bti’s efficiency in killing larvae. Results obtained from Fort Ternan show a similar pattern, where early instars were recorded from
Bti intervention habitats while the late instars were fewer within habitats provided with
Bti, predatory fish and those under drainage. The efficacy of
Bti in reducing mosquito larval populations recorded in this study is comparable to the studies by Majambere et al. [
36] and Fillinger et al. [
37]. No residual effect of
Bti was observed during the post intervention period as both early and late instar anopheline and Culex larvae were present within habitats previously applied with
Bti.
The use of predatory fish for mosquito control has not been widely used in Africa; however, large scale trials using various species of larvivorous fish to control anopheline and culicine larvae have been reported in the Mediterranean region [
42].
Gambusia affinis has been in use over a long period of time for mosquito control in countries such as Afghanistan, Cyprus, Egypt, Sudan and Jordan. Tilapia (
Oreochromis) and
Aphanius dispar are other species that are in use for mosquito control [
42]. In western Kenya, Howard et al. [
43] have shown the potential of
Oreochromis niloticus, for mosquito control. In the current study when predatory fish
, G. affinis was used as a single option in erosion pits, few early instars were recorded whereas no late instars were recorded from the same habitats. The results indicate that the female mosquitoes would still oviposit within habitats with fish but the larvae are fed upon before reaching their late stages.
In Lunyerere, manipulation of breeding habitats through shade provision by growing Napier grass [
44] or arrow roots grown along drainage canals are capable and promising in the control of anopheline larvae. Source reduction involved modification of the existing canals to increase water flow so that larvae would be flushed out into a fast moving water canal/ river where they eventually die. No anopheline and culicine larvae were recorded from drained habitats in Lunyerere and Fort Ternan. If well implemented, drainage is successful; however the open drains or drainage canals need to be maintained to remove any debris that may slow water flow and lead to creation of pools of stagnant water that provide mosquito grounds. A number of studies have reported the successes of malaria reduction and eradication through environmental management projects [
45‐
47]. However, such strategies have not been fully exploited in areas where malaria still remains a major health problem such as in western Kenya.
The LSM strategies were executed differently in the two study sites because of the variation in the nature of the breeding habitats. Lunyerere being a reclaimed swamp area, drainage canals were made for land reclamation to give way for farming. Failure of maintenance of the drainage canals led to creation of stagnant pools of water that are preferred mosquito breeding habitats. The findings show that source reduction through drainage and habitat manipulation compared well with larviciding. Arrow roots do well in swampy areas and because they are a good source of carbohydrates, they ensure food security for the land owner/farmer, in-addition, they may also be sold to generate income. On the other hand in Fort Ternan, stagnant water resulting from leakage of water pipes/taps and pools of water on the fringes of Kipchorian river were drained, whereas erosion pits formed pond like habitats that were stocked with predatory fish. The alternatives strategies applied in both sites compared well with the application of
Bti. For a reclaimed swamp area environmental management would work well while in areas with pond-like habitats the use of predatory fish would be beneficial. However, a number of limitations were experienced during the study. The results for Fort Ternan should be interpreted carefully because in this area, breeding habitats were fewer when compared to Lunyerere, thus limiting the number of replicates. Previous studies in the same area found the densities of immature malaria vectors to be very low [
7]. The sampling period coincided with the dry period (November to February) and a number of potential breeding habitats dried out further reducing the number of replicates. Thus Fort Ternan might not have been a good site for such as trial when compared to Lunyerere where the vector breeding was throughout due to the presence of favourable breeding grounds. Nevertheless, the results are promising and future studies should be done in an area where vectors breed in more habitats to allow for a better judgement of the strategies applied.
Recent times have witnessed the successes of an integrated approach of malaria control in many countries through the combination of ITNs and IRS as advocated by the Roll Back Malaria initiative [
25]. Although insecticide treated bednets and indoor residual spray are highly effective for the control of indoor biting and resting mosquitoes, due to vector avoidance and possible behaviour changes [
29,
30,
48] strategies targeting outdoor vectors are required to complement existing measures. In a previous study, it was found that the community members were willing to take part in larval source reduction but they lacked evidence-based results on control strategies that can be used with locally available resources [
49]. As a step towards achieving this goal, the manuscript provides results of a small scale field trial on LSM strategies that can be incorporated into an IVM program, using locally available resources in comparison to the application of larvicides and the use of predatory fish for mosquito larval control.
Competing interests
The authors declare that there is no conflict of interest.
Authors’ contributions
This study was conceived by Imbahale SS and Takken W. Imbahale SS supervised field data collection, did data analysis and drafted the manuscript. Githeko AK, Mukabana WR and Takken W assisted with study design and logistical issues. All authors have read and approved the final version of manuscript.