Entomological survey in two communes with residual malaria transmission in Gia Lai Province in the central highlands of Vietnam

In 2018, the National Malaria Control Programme in Vietnam switched from prioritizing malaria control to elimination. However, with the ongoing elimination programme, there are still areas where residual malaria transmission persists, including the central highlands. This entomological survey was conducted to evaluate Anopheles diversity and host-seeking activity of Anopheles vectors in two communes with very low malaria transmission in Gia Lai Province. Anopheles species were collected in Ia DReh commune and Ia KDam commune, Gia Lai Province in the central highlands of Vietnam. Collections were conducted using human-baited double net trap, light trap and manual aspiration collections around cattle sheds, in the dry and rainy season. Mosquito specimens were identified morphologically, and members of species complexes were distinguished molecularly. Mosquito night-feeding patterns were investigated during the dry and rainy seasons. Overall, 18,835 specimens including 19 taxa were collected in Ia KDam and Ia DReh communes. These included the primary malaria vectors, Anopheles dirus and Anopheles minimus, and other secondary vector species. Anopheles dirus was observed to be an anthropophilic species, whereas An. minimus and a number of secondary vectors were observed to be zoophilic. Anopheles vagus was the dominant species, followed by Anopheles sinensis and Anopheles peditaeniatus. The majority of specimens were collected in the rainy season due to the relatively large number of An. vagus, while An. peditaeniatus, An. dirus, Anopheles kochi, Anopheles monstrosus and Anopheles tessellatus were collected in greater numbers during the dry season. The peak of host-seeking activity for An. dirus, An. sinensis, and An. vagus was between 18.00 and 19.00 h. This study provided information on the diversity, seasonal prevalence and behaviour of Anopheles at the study sites. Identifying the diverse mosquito fauna in the central highlands of Vietnam allows species-specific control measures to be implemented by the National Programme to reduce malaria in areas of very low malaria transmission. The peak Anopheles host-seeking activity observed in this study was between 18.00 and 23.00 h, which highlights the need to better characterize Anopheles behaviour in this region of Vietnam to inform on vector control strategies.

of insecticide-treated nets (ITNs), long-lasting insecticide nets (LLINs), social mobilization, targeted residual indoor spraying, and improved anti-malarial drug treatment. These strategies significantly reduced the spread of malaria in many areas in the country [2][3][4]. However, despite these measures, there remain residual malaria transmission areas, predominately in forested regions of central and southern Vietnam [5][6][7][8][9][10]. These very low malaria transmission areas harbour primary malaria vectors and there are reports that long-lasting insecticidal net (LLIN) use is less common [11,12]. In deforested rural areas, where land has been cleared for cultivation, many secondary malaria vectors are present and may contribute to maintaining malaria transmission outside of forested areas [10].
Correctly identifying the mosquito fauna in malariaendemic areas is critical to implementing species-specific control measures and other public health interventions. The identification of mosquito vectors is mainly based on morphological diagnostic characters. However, misidentifications of closely related species, with overlapping morphological characters, is possible and is more likely amongst sibling vector species [17]. In these circumstances, molecular identification has been used for definitive identification [18][19][20][21].
In Gia Lai Province, in the central highlands of Vietnam, there are many individuals who depend on the forest, forest fringe and deforested areas for subsistence (agriculture and wood cutting activities) with potential exposure to malaria vectors [22]. In 2020, Krong Pa was the most impacted district with 459 malaria cases, while 43 cases were reported in Ia Pa District (Gia Lai Province Center for Disease Control and Prevention-CDC pers. comm.). A survey of Anopheles vectors was conducted in 2019 to inform vector control strategies in residual malaria transmission areas of Gia Lia Province. The study focused on characterizing diversity and host-seeking activity of Anopheles vectors in Gia Lai Province.

Study sites
Gia Lai is a mountainous province in the northern region of the central highlands of Vietnam (Fig. 1), and has a tropical monsoon climate, and two distinctive seasons: the rainy season (May to October) and the dry season (November to April). The average annual rainfall is 2100 mm to 2200 mm, the average temperature is 22 °C to 25 °C and the average annual air humidity is about 80% (Gia Lai Province Local Health Office, pers. comm.).
Entomological surveys were conducted during the rainy and dry seasons in Ia DReh (IDR) commune, Krong Pa District, and in Ia KDam (IKD) commune, Ia Pa District, Gia Lai Province (Fig. 1). Village sites (inside the villages) are those where people live. Village sites contain houses and small gardens with few trees; whereas farm (mixed with cassava, sugar cane and watermelon plantations) and forest sites (tropical dense forest in IDR and deforested area in IKD) are places where villagers conduct agriculture and forestry activities, but are not permanent residences (TQN pers. obsv.). IDR commune is 110-140 m above sea level and includes 8 villages: Hvut, Nai, Drai, Hyu, Djrong, Tring, Bau, and Chum Kia. There is a nature reserve named Ea So, which is comprised of dense tropical forest approximately 10 km from the villages. IKD commune is located 232 m above sea level, and includes deforested and open areas used for farming purposes. This commune comprises 7 villages: Hbel, Plei Toan 1, Plei Toan 2, Plei KDam 1, Plei Kdam 2, Bau, and Chroh Ko. In 2020, the population was 5,363 people in IDR (2642 women, 2721 men) and 4176 people in IKD (2341 women, 1835 men). In both communes, 30% of population were between 15 and 49 years old. In the last years, the strategies of vector control have been the use of LLINs and insecticide-treated nets (ITNs), where IDR has been supplied with LLINs, and IKD with LLINs and ITNs (Gia Lai Province, CDC pers. comm.).
In 2019, 103 and 131 symptomatic malaria cases confirmed by blood film microscopy were reported in IDR and IKD communes, respectively (Gia Lai Province, CDC pers. comm.). The annual parasite incidence was 21.2 (103/4.857) cases per 1000 populations for the IDR commune and 35.2 (131/3.717) cases per 1000 populations for the IKD commune. Based on the annual parasite incidence < 100 cases per 1000 populations, both communes would be classified as areas of very low malaria transmission [23].

Anopheles collection
Entomology surveys were conducted in the dry season (1-28 January, 2019) and in the rainy season (29 June-24 July, 2019). During each survey, mosquitoes were collected in both communes, IDR and IKD, using light traps (LT) from 18.00 to 06.00 h, human-baited double net traps (HDN) [24], and using manual aspirations around cattle sheds (CS) from 18.00 to 24.00 h. HDN and CS collections were conducted by one collector (i.e., one collector per trap/method).
For surveys in villages sites, during each season, one indoor HDN and one outdoor HDN was set up for 18 nights; CS collections were conducted for 24 nights; and LT collections were set up indoors (n = 2) and outdoors (n = 2) for 18 nights in IDR and IKD. For surveys in farm and forest sites, one outdoor HDN was set up for 10 nights at IDR and IKD during each season; CS collections were conducted for two nights during the rainy season in IDR, whereas in IKD were conducted for three nights during the dry season and for four nights during the rainy season; no LT was set up in IDR, whereas two LTs were set up in IKD for five nights only during the rainy season.

Anopheles identification
Mosquitoes were morphologically identified under stereomicroscopes using keys developed by the National Institute of Malariology, Parasitology and Entomology [25]. Molecular methods were used for the identification of species complex/group members, except Barbirostris Complex was not identified molecularly, it was identified by morphological characters as An. barbirostris sensu lato (s.l.). DNA from head and prothorax of collected mosquito specimens was extracted using the QIAamp DNA Mini Kit-DNA Purification from Tissues as described in the manufacturer's instructions (Qiagen, Hilden, Germany). To identify species complex or group of Anopheles, the molecular methods used are those described by Garros [20]. Specimen identification was recognized according to the length of the product for each polymerase chain reaction (PCR)-based identification (

Statistical analysis
Considering the heterogeneous number of traps and number of collections, specific comparisons of each collection method between communes, and according to the season were employed. For statistical analysis, the Monte Carlo Method, or Fisher's Exact Test were used when the sample size was small. These statistical tests were conducted using IBM SPSS statistics, version 25, with statistical significance defined as a p < 0.05.

Density and comparison of mosquitoes according to each collection method
Anopheles vagus was collected in high density using CS (85.0%, 15,193  distribution, due to animal movements, and the bad weather impaired the collection of mosquitoes using LT in farm/forest sites, therefore no comparison was made between farm and forest sites. Additional file 1 reported the number of Anopheles mosquitoes collected outdoor and indoor using HDN and LT.

Mosquitoes collected in village sites using HDN, CS, and LT
In village sites, in both IDR and IKD communes, species differed depending on the season, except An. sinensis (49 and 51% collected in dry and rainy seasons, respectively, in IKD) ( Table 6). Anopheles aconitus, An. kochi, An.   (Table 6). Total Anopheles species collected in village sites were more common in the rainy season due to the high density of An. vagus (Table 6). When An. vagus was excluded, and comparisons were made between the other Anopheles mosquitoes, they are more abundant in the dry season at the two communes (82.5% in IDR, 1,130/1,370, and 69.2% in IKD, 662/957) (p < 0.05, Monte Carlo Method) ( Table 6).

Mosquitoes collected in farm/forest sites using HDN
In farm and forest sites, An. philippinensis were collected only in the dry season in IDR, and An. aconitus in IKD; while Anopheles barbirostris s.l., An. maculatus, An. peditaeniatus, and An. vagus were collected only in the rainy season in IDR, and An. minimus in IKD (  (Table 7). In IDR, Anopheles mosquitoes were more abundant in the dry season (67.6%, 100/148), while in IKD they were slightly more common in the rainy season (53.1%, 60/113) (p > 0.05, Fisher Exact Test) ( Table 7). High density of An. dirus was captured using HDN in farm/forest sites in both dry and rainy season, following by An. tessellatus in the dry season and An. sawadwongporni in the rainy season. Anopheles vagus was found in both dry and rainy season at village, and farm and forest sites (Table 8).

Observation on host-seeking activity of malaria vectors in humans (HDN) and cattle (CS)
The observation of host-seeking activity for primary and secondary malaria vectors in seeking humans and cattle are shown in the Additional files 2 and 3. Malaria vectors collected in high density are presented in Figs. 2 and 3. In IDR commune, the peak of host-seeking activity for An. dirus collected using HDN in farm and forest sites occurred between 18.00 and 19.00 in the dry season ( Fig. 2; Additional file 2). For mosquito species collected around CS in village sites, the peak occurred between 19.00 and 21.00 for An. vagus, between 19.00 and 22.00 for An. sinensis, and between 21.00 and 22.00 for An. peditaeniatus in the rainy season; in the dry season, the peak  of host-seeking activity for An. sinensis and An. vagus occurred between 18.00 and 19.00, and between 20.00 and 21.00 for An. peditaeniatus ( Fig. 2; Additional file 2). The peak of host-seeking for An. vagus collected in farm and forest sites during the rainy season was between 19.00 and 20.00 ( Fig. 2; Additional file 2. In village sites in IKD commune, the peak of host-seeking activity was between 20.00 and 21.00 for An. sinensis and An. vagus during the rainy season; and between 19.00 and 20.00 for An. An. sinensis and An. vagus, and between 20.00 and 21.00 for An. peditaeniatus in the dry season ( Fig. 3; Additional file 3). In farm and forest sites of IKD, the peak of host-seeking activity was between 19.00 and 20.00 for An. barbirostris s.l., between 20.00 and 21.00 for An. vagus, between 21.00 and 22.00 for An. aconitus and An. minimus, and between 22.00 and 23.00 for An. sawadwongporni in the rainy season. In the dry season, the peak for An. vagus occurred between 18.00 and 19.00, and for An. sawadwongporni between 19.00 and 20.00 in the dry season ( Fig. 3; Additional file 3).

Anopheles diversity and seasonality
Herein, the two primary malaria vectors, An. dirus and An. minimus were found at the two study sites of very low malaria transmission in Gia Lai Province, along with 9 secondary malaria vectors: An. aconitus, An. barbirostris s.l., An. harrisoni, An. maculatus, An. peditaeniatus, An. philippinensis, An. sawadwongporni, An. sinensis, and An. vagus. Overall, 19 Anopheles species (14 taxa in IDR, and 18 taxa in IKD) were identified at the two study sites. From the same province, in Kong Chro district (40 km from IDR and 70 km from IKD), Garros et al. [28] collected 11 Anopheles species, which included An. dirus s.l., An. maculatus s.l., and An. minimus s.l.. Identification of these mosquito specimens were based on morphology only, and thus it was not possible to identify which members of each species complex were found in their study. All Anopheles species found from the study by Garros et al. [28] were collected from the IKD and IDR sites in this study, except An. jeyporiensis, and An. nivipes. By combining the Anopheles species identified by Garros et al. [28] and those collected herein, a total of 21 Anopheles species have been identified from Gia Lai Province.
Anopheles diversity was greater in the forest sites (12 taxa in village sites, and 18 taxa in forest sites). Anopheles dirus and An. minimus, and the secondary malaria vectors, An. aconitus, An. barbirostris s.l., and An. sawadwongporni were more common in farm/forest sites. Although previous studies have incriminated An. aconitus and An. sawadwongporni, and other Anopheles species (An. harrisoni, An. maculatus, An. pampanai, An. peditaeniatus An. philippinensis, An. sinensis) in maintaining malaria transmission in rural areas of the country [10,29], still little is known about the role of secondary vectors in residual malaria transmission in rural areas.
Anopheles dirus and An. minimus were found only in forest sites. This might indicate the prevalence of 'external' malaria infection (infection outside the village) [8] observed in workers associated with forest sites in IDR (IDR health station, pers. comm.). Rainforests and secondary forests or plantations (rubber, fruits) provide appropriate breeding sites for An. dirus [30,31]. The farm and forest sites in both communes are around 3 km and 10 km (IKD and IDR, respectively) away from the villages, and outside the flight range (approximately 1.5 km) for An. dirus [30]. Therefore, the long distance between the breeding sites and the villages may be the reason for the absence of An. dirus in village sites. The breeding sites for An. minimus are usually ubiquitous [32], but perhaps in IKD this species is a habitat specialist, and thus found only in the farm/forest sites. Also, these sites can offer more host variety (i.e., cattle, chicken, dog). However, ecological studies are needed to clarify the ecology of An. minimus in Gia Lai Province.
Anopheles maculatus and An. sawadwongporni were the only vectors molecularly identified in the Maculatus Group. Anopheles sawadwongporni was the most collected species of the Maculatus Group in this study, making up 1.3% (240/18,835) of all collected species, and only a small number of An. maculatus (IDR, n = 1; IKD, n = 5)  Due to its ability to survive in small pools of turbid and organically polluted water [34], An. vagus has been found in high densities in countries of Southeast Asia (SEA) [8,10,28], including An. vagus positive for circumsporozoite protein (CSP) in mosquitoes in Bangladesh [35,36]. However, there is no published information about the vector status of this species in Vietnam. Anopheles peditaeniatus have been found to be positive for CSP in Indonesia [37] and Thailand [38], and An. sinensis was involved with malaria transmission in China [39] and epidemics in Korea [40]. Anopheles peditaeniatus and An. sinensis were found to be positive for CPS in Quang Binh Province, Vietnam [10], with An. sinensis considered the major malaria vector for the province. The adaptable An. vagus has been observed in different environments [10,34] and because of its opportunistic and zoophilic behaviour [8,10] limited research has been carried out on this species. Anopheles vagus appears not to play an important role in malaria transmission at the two study sites. A high abundance of mosquitoes is expected in the rainy season, where more breeding sites are available and conditions are favourable for mosquito development. Indeed, more mosquitoes were collected during the rainy season compared to the dry season in this study, similar to reports from Thailand and Bangladesh [30,41]. Similarly, an increased abundance of An. dirus and An. minimus has been observed in Vietnam during the rainy season between September and November [42]. Herein, An. minimus was present only during the rainy season, however, when combining the collections from the village and farm/forest areas of the two sites (IDR and IKD) most of the An. dirus (74.4%, 67/90) were collected in the dry season. Rain was especially intense in the forest sites in both communes during the rainy season, and apparently flooded the breeding sites of An. dirus and other species that have more suitable habitats in the forest in IDR. On the other hand, An. minimus and the dominant An. vagus were more common during the rainy season, probably due to the temporary breeding sites being more available in IKD for their development. Nevertheless, An. dirus was still found in the rainy season, as well as secondary malaria vectors collected throughout both seasons. This suggests that favourable conditions for vector development exist year-round, which ensures the maintenance of malaria parasites. Consequently, the epidemiologic triad of host (human, animal), agent (Plasmodium + Anopheles), and environment conditions contribute to the maintenance of malaria transmission in the study areas.

Collection and behaviour of malaria vectors
Due to logistical limitations, the distribution of traps in the two communes and the number of collections were not homogeneous. The limitations included difficulty in setting up traps, arranging and maintaining cows in farm and forest sites, and loss of LTs. Unfortunately, many traps were knocked over and found empty, probably because of animal movement during the night. Also, during the collection, the weather impacted sample collection, with heavy rain and flooding which made it difficult to collect the mosquitoes at the communes. Nonetheless, specific comparisons of the mosquitoes collected for each collection method was carried out.
Studies of malaria vectors in SEA have used a variety of methods to collect mosquitoes, including human baits (human landing catches-HLC and HDN) [8,10,15,24,43], animal baits (buffalo, cattle) [8,10,15,33], and LTs [33,[44][45][46] to collect mosquitoes. Light traps are a costeffective method for mosquito collection; however, the efficacy varies by location [45,47]. Another cost-effective method for the collection of vector species is HDN, an alternative to HLC, which eliminates exposing collectors to mosquito bites [24]. Herein, a high density and diversity of mosquitoes were collected using CS. In other studies, similar results were reported using CS, with a high diversity of mosquitoes collected including primary and secondary malaria vectors [10,28,48]. Despite the simultaneous collection using human (HDN) and cattle bait (CS), it was not possible to calculate the anthropophilic index due to the heterogeneous distribution of traps and number of collections, and the prominent zoophilic behaviour observed of all Anopheles species, except An. dirus.
Anopheles dirus was observed to be the most anthropophilic species in this study and is consistent with other studies conducted in SEA [13,30,31,33,49]. However, there are reports of An. dirus with strong zoophilic behaviour in Kanchanaburi Province, Thailand [50] suggesting some variance in behaviour across SEA. Anopheles minimus, together with many secondary vectors identified in this study, showed zoophilic behaviour, which has been similarly observed in other regions in Vietnam and neighbouring countries [10,15,[51][52][53][54]. Anopheles minimus host-seeking behaviour is similarly variable with anthropophilic behaviour observed in Khanh Hoa Province (Vietnam) and Vientiane Province (Laos) while Van Bortel et al. [32] reported heterogeneity in host choice. There are reports An. minimus are attracted to cattle when abundant, but are attracted to humans when cattle are scarce [15,32,55]. In this study An. minimus was found exclusively in farm/forest areas in IKD, where there were many types of host, such as cattle, chicken and dogs. The close proximity of cattle to household areas has been associated with an increase in malaria prevalence [56,57] suggesting a relationship to host-seeking behaviour and horticulture practices. A future study of blood meals in An. minimus is required to better understand the impact of host choice on vector survival and reproduction in areas of malaria transmission. Do Manh et al. [10] conducted a malaria survey in Quang Binh Province (Vietnam) using paired collections of buffalo and humans (with 15 m of distance between them), and observed that 97.4% (4,489/4,610) of the anopheline collected were attracted to buffalo and only 2.6% (121/4,610) to humans using HLC. The authors suggested that animals (buffalo/cattle) around human settlements may play a zooprophylactic role [10]. In this study, it was observed that anopheline were more attracted to animals than human, similar to Do Manh et al. findings [10]. To reduce vector density and longevity, Do Manh et al. [10] proposed a strategy of maintaining animal baits (cattle, buffalo) around human habitation and implementing barrier spraying to supplement other intervention methods such as IRS and the use of LLIN.
The early host-seeking time of 18.00 to 19.00 observed for An. dirus, An. sinensis and An. vagus coincided with human activities outdoors. This study implies that the use of LLIN after 21.00 will have limited impact on reducing Anopheles bites. Villagers and forest workers usually go to bed after 21.00 (TQN pers. obsv.), therefore, the use of LLIN in areas where vectors bite outdoors and/or early in the evening when people are still active will not be very effective [58]. Additionally, some bed nets appeared worn with large holes or gashes, limiting their ability to protect users. LLINs are a critical component of the NMCP, however, there are reports of limited use despite nearuniversal availability [11,12,22]. Perhaps, an alternative to support the strategies to prevent malaria transmission could include the distribution of repellent or repellent clothing (treated coveralls) to the villagers-workers in addition to the LLINs as cost-effective methods for protection against malaria vectors. Moreover, additional research based on practical personal protection tools (i.e., those do not require behaviour change and are easy to use) is required to eliminate residual malaria transmission [5] in these areas.

Conclusions
Nineteen Anopheles species were identified at the two study sites in Gia Lai Province, including two primary malaria vectors, An. dirus and An. minimus and 9 secondary malaria vectors. During the course of this study, An. dirus displayed anthropophilic preferences, whereas An. minimus and a number of the secondary vectors displayed zoophilic preferences for host. The observed results provided information on the diversity, seasonal prevalence and behaviour of Anopheles in the study sites. Moreover, the peak (18.00 to 23.00) Anopheles host-seeking activity was observed in this study highlighting the future need to focus on Anopheles behaviour in this region of the central highlands of Vietnam.