This nationwide entomological survey documents the bionomics of malaria vectors in ten sites of Laos where malaria elimination is planned for 2030. This is the largest entomological study implemented in the country since the pioneer work of the MALVECASIA project in the 2000’s (Obsomer et al., 2007; VanBortel et al.,2008). Our data are timely and of great importance to guide decision making in vector control in the country. Indeed, accurate information on vector composition, bionomic and distribution (both spatial and temporal), are keys to design and implement scalable and locally adapted vector control interventions.
Results showed a great diversity of Anopheles species in the study areas with 25 different species/complexes morphologically identified (Supplementary Table 1). All the Anopheles complex species collected were already described in Laos (Pholsena et al., 1992; Meek et al., 1995; Kobayashi et al., 1997; Toma et al., 2003; Wythilingam et al., 2003; Motoki et al., 2019) but thirteen Anopheles sp. from three different groups (i.e. Funestus, Leucosphyrus and, Maculatus) could be identified for the first time for some species, using molecular tools. These findings provide additional information to the checklist of the 42 Anopheles species of Laos recently updated by Motoki et al. (2019). Moreover, this provides us important data on the relative proportions of primary versus secondary vectors within the three groups mentioned above. For example, the three primary vector species, identified by qPCR, An. dirus s.s., An. maculatus s.s. and, An. minimus s.s. represented 89, 36 and, 33 percent of the total mosquitoes of the Dirus and Minimus complexes and Maculatus group, respectively. In the Maculatus group, An. rampae represented 34 percent of the total but this zoophilic species is not a malaria vector (Sinka et al., 2011). In contrast, within the same group An. sawadwongporni represented 14 percent of the total and is considered as a very efficient vector in Thailand (Marasri et al., 2017). Within the Minimus complex, An. aconitus accounted for 57 percent of the total and represented more than 10 percent of the total number of the mosquitoes collected. This species is highly zoophilic and exophagic (Junkum et al., 2007; Gould et al., 1967), and Manh et al. (2010) showed it was to some extent, responsible to maintain transmission in rural communities, and deforested areas in north-central Vietnam. The abundance of other secondary malaria vectors was relatively high, such as An. nivipes (19% of all Anopheles sp. collected), An. philippinensis (4.8%) and, to a lesser extent An. barbirostris (1.9%). All these species are mostly zoophilic but they can also bite humans. Several studies implemented in Laos (Kobayashi et al., 1997, 2000; Toma et al., 2002) showed that An. philippinensis and An. nivipes are able to bite both human and animals and the authors suspected them to be responsible for malaria transmission in paddy field areas in Khammouane province. Indeed, both of these species were previously found infected by P. falciparum or P. vivax in Laos and in other GMS countries (Harbach et al., 1987; Toma et al., 2002; Rattanarithikul et al., 2006). Clearly more work has to be done to determine the behavior and ecology of secondary vectors and their role in transmission.
The abundance and diversity of the Anopheles mosquitoes showed a seasonal trend. As expected, there were more Anopheles mosquitoes collected during the rainy season (68%) than during the dry season (32%) and the number of mosquitoes collected varied significantly according to the location. For example, more than 3,500 Anopheles mosquitoes were collected in Vientiane province alone, against only 150 specimens in Savannakhet, but overall, these results showed that people in rural Lao villages are constantly exposed to malaria vector mosquitoes throughout the year. Furthermore, the study on the mosquito biting preference showed that even if they are highly zoophilic, primary and secondary vectors are biting humans constantly during the night, both indoors and outdoors. Previous studies showed that there was a direct link between close proximity between human and cattle, biting rates and malaria prevalence (i.e. zoopotentiation; Donnelly et al., 2015; Bouma and Rowland 1995). In most of our study sites, cattle and livestock were largely present around the houses from dusk to dawn and the same species were found biting both human and cows but in higher proportion on this latter. Every day before sunset, the cattle owners of the villages bring back their animals to their yard and sometimes under their traditional wooden houses or in a dedicated place in the village, thus increasing the risk for villagers living near these animals to being bitten outdoors. Almost thirty percent of primary and secondary vectors were collected outdoors before 10:00PM or after 5:00AM when people are still outside. This shows the importance of personal protection and other outdoor related control measures, such as zooprophylaxis, to tackle malaria transmission in these remote areas. In the systematic review of Donnelly et al. (2015), the authors pointed out that zooprophylaxis may be part of Integrated Vector Management (IVM) in areas where the dominant vectors are highly zoophilic and the livestock are kept away from human sleeping quarters. However, their results also showed that when vector preference is opportunist, varied or unknown, there are no evidence to support the use of zooprophylaxis and zoopotentiation could even be increased. In Laos, research regarding this method is clearly needed to validate its usefulness within the large diversity of environments, vectors characteristics and socio-economic factors.
In total, sixty-three percent of the vector were collected outdoors which is in adequacy with the results of Chaumeau et al. (2018) who estimated that sixty-five percent of the potential infective bites are not prevented by mosquito bed nets because of outdoor and early biters. Twenty-two percent of the malaria vectors were collected in the villages indoors between 10:00 PM and 5:00 AM when the people are sleeping. Although this represent a relatively low percentage, it highlights the crucial need to provide household with bed nets to protect people during this specific period of the night. Kobayashi et al. (2004) already confirmed in the 2000’s, the efficacy of treated bed nets in highly endemic areas of Laos and since the country-wide distribution programs, implemented by the Ministry of Health, of first ITNs and then LLINs, malaria prevalence has dramatically decreased (WHO 2016). Vector control with the use of pyrethroid insecticides in Laos is currently possible as the primary and secondary vectors are still mostly susceptible (Marcombe et al., 2017). However, continuous monitoring and the use of different insecticide family is recommended as resistance is likely to evolve in some parts of the country using high amount of insecticides for pest control. In a recent study, Souris et al. (2017) highlighted the space-time distribution of the environmental risk of Anopheles presence, potential insecticide emergence, insecticide resistance, and risk of exposure to these threats for the human population in Laos. Their results showed that the probability of insecticide resistance in malaria vectors is greater in the southern part of the country, specifically in Champasak and Attapeu provinces, bordering Cambodia, Thailand and Vietnam. In these areas, malaria incidence is among the highest and the resurgence have been attributed to large-scale population movements (both within Lao and across national borders) as well as forest-related economic activities (Khounnavong et al., 2017).
Malaria transmission in the Mekong region is currently concentrated in forested and rural areas and along national borders from where the disease is likely to spread to other areas due to the movement of population groups (WHO 2019, Kounnavong et al., 2017; Inthavong et al., 2017). A part of the transmission is taking place outdoors in the villages as previously mentioned, but the remaining transmission occurs outside of the villages, especially in the forest. In Laos, significant correlations between working and sleeping habits in the forest and malaria incidence were reported (Inthavong et al., 2017; Phommasone et al., 2016). In forested areas, An. dirus is thought to be the dominant malaria vector of the southern parts of the country (Sidavong et al., 2004; Toma et al., 2002; Vythilingam et al., 2003; Kobayashi et al., 2004). However, very few specimens were collected in our study (n = 43) and this could be explained by its specific breeding sites usually being located in forested areas nearby the villages (Souris et al., 2017, Hii et al., 2013). Mosquito collections conducted in forested areas of the Lao-Thailand border at the same period of time, showed that An. dirus was the predominant species biting humans (Marasri et al., 2017). This study showed that during the rainy season, human biting rates of An. dirus (HBR = 0.91) were more than thirty times higher than within village locations (HBR = 0.03). We highly recommend to conduct mosquito collections in the southern forested and remote areas of Laos to better understand the role of An. dirus in malaria residual transmission among people working in plantations and forest camps.
Our study on malaria parasites infections in the mosquitoes collected could not give us valid indications on the malaria transmission intensity and patterns in our study sites. Indeed, more than 4,000 mosquitoes were tested for Plasmodium incidence and only two specimens were positives; one Anopheles aconitus specimen from Phongsaly province and one An. minimus specimen from Vientiane Province were positive with P. falciparum, with a mean sporozoite rate of 0.04%. This confirmed the recent work of Sumarnrote et al. (2018) where no infected mosquitoes could be found in the Kanchanburi province in Thailand during a malaria outbreak. This also confirms previous studies implemented in the early 2000’s in Laos, when malaria transmission was even higher than nowadays (WHO 2016, Toma et al., 2002; Kobayachi et al., 2004; Vythilingam et al., 2003), sporozoite rates in mosquitoes were very low. These results can be explained by i) the low number of night collections conducted during the dry and rainy seasons hence limiting the chance to catch Plasmodium-positive mosquitoes in low transmission settings and ii) by the strong zoophagic preferences of both primary and secondary malaria vectors hence limiting the human–vector exposure.
Currently, research on alternative strategy for vector control is nonexistent in Laos. It is imperative to determine joint research priority axis in Laos and in the GMS with regards to additional vector control tools (VCTs) that could complement insecticide-treated nets (ITNs) and indoor residual spraying (IRS) to achieve malaria elimination (Williams et al., 2018). VCTs should take into account the dynamics of the transmission, as well as the ecology of malaria vectors in local settings. For example, veterinary approaches such as the use of insecticide-treated mosquito nets fenced around cattle (Maia et al., 2012) or the use of endectocides by injection in livestock (Chaccour et al., 2016) may be interesting strategies to target the zoophilic and exophagic zoophagic malaria vectors (e.g. An. maculatus, An. minimus and An. sawadwongporni). The use of mosquito-proofed housing could be useful to protect people from endophagic mosquitoes such as Anopheles dirus, An. nivipes, An. barbirostris, and An. Philippinensis. Inthavong et al. (2017) clearly showed that households in villages with high malaria incidence were significantly more likely to have an open space on the house surface compared to villages with low incidence.