Resting behaviour of malaria vectors in a highland and a lowland site of western Kenya: Implication on malaria vector control measures

Background Understanding the interactions between increased insecticide resistance in field malaria vector populations and the subsequent resting behaviour patterns is important for planning adequate vector control measures in a specific context and sustaining the current vector interventions. The aim of this study was to investigate the resting behavior, host preference and infection with Plasmodium falciparum sporozoites by malaria vectors in different ecological settings of western Kenya with different levels of insecticide resistance. Methods Indoor and outdoor resting Anopheline mosquitoes were sampled during the dry and rainy seasons in Kisian (lowland site) and Bungoma (highland site), both in western Kenya. WHO tube bioassay was used to determine levels of phenotypic resistance of first generation offspring (F1 progeny) of malaria vectors resting indoors and outdoors to deltamethrin. PCR-based molecular diagnostics were used for mosquito speciation, genotype for resistance mutations and to determine specific host blood meal origins. Enzyme-linked Immunosorbent Assay (ELISA) was used to determine mosquito sporozoite infections. Results Overall, 3,566 female Anopheles mosquitoes were collected with Anopheles gambiae s.l [In Bungoma, An. gambiae s.s (90.9%), An arabiensis (7.6%) and in Kisian, An. gambiae s.s (38.9%), An. arabiensis (60.2%)] being the most abundant species (74.7%) followed by An. funestus s.l (25.3%). The majority of An. gambiae s.l (85.4 and 58%) and An. funestus (96.6 and 91.1%) were caught resting indoors in Bungoma and Kisian respectively.Vgsc-1014S was observed at a slightly higher frequency in An. gambiae s.s hereafter(An. gambiae) resting indoor than outdoor (89.7 vs 84.6% and 71.5 vs 61.1%) in Bungoma and Kisian respectively. For An. arabiensis, Vgsc-1014S was 18.2% indoor and outdoor (17.9%) in Kisian. In Bungoma, the Vgsc-1014S was only detected in An. arabiensis resting indoors with a frequency of 10%. The Vgsc-1014F mutation was only present in An. gambiae resting indoors from both sites, but at very low frequencies in Kisian compared to Bungoma (0.8 and 9.2% respectively. In Bungoma, the sporozoite rates for An. funestus, An. gambiae, and An. arabiensis resting indoors were 10.9, 7.6 and 3.4 % respectively. For outdoor resting, An. gambiae and An. arabiensis in Bungoma, the sporozoite rates were 4.7 and 2.9 % respectively.Overall, in Bungoma, the sporozoite rate for indoor resting mosquitoes was 8.6% and 4.2% for outdoors. In Kisian the sporozoite rate was 0.9% for indoor resting An. gambiae. None of the outdoor collected mosquitoes in Kisian tested positive for sporozoite infections. Conclusion The study reports high densities of insecticide-resistant An. gambiae and An. funestus resting indoors and the persistence of malaria transmission indoors with high entomological inoculation rates (EIR) regardless of the use of Long-lasting insecticidal nets (LLINs). These findings underline the difficulties of controlling malaria vectors resting and biting indoors using the current interventions. Supplemental vector control tools and implementation of sustainable insecticide resistance management strategies are needed in western Kenya.

Malaria still remains a major public health concern in sub-Saharan Africa, responsible for an 78 estimated 219 million cases and 435,000 deaths despite the massive investments in scaling-up 79 indoor anti-vector interventions [1]. Remarkable advances in the fight against malaria have been 80 achieved within the past decade mainly through the massive scale-up of long-lasting insecticide-81 treated nets (LLINs) and indoor residual spraying (IRS) in many localities [2,3]. Despite the 82 increased efforts, it is worrying that no significant progress has been made in reducing global 83 malaria cases in the year 2015-2017 period [1], with some regions in sub-Saharan Africa, 84 previously reported to experience a resurgence of malaria including western Kenya [4]. This 85 transmission recurrence is partly attributed to the emergence of insecticide resistance and 86 behavioural modification that have arisen as an adaptation by mosquitoes in response to high use 87 of insecticides for vector control [5,6]. All these factors have the potential to weaken malaria 88 control programs thus posing a serious threat in the fight against malaria [7]. 89 The current vector control interventions take advantage of susceptible mosquito behaviors. These 90 interventions are based on the observation that malaria vectors prefer to bite humans indoors late 91 at night and often rest inside houses after blood feeding hence, they will be exposed to sufficient 92 levels of insecticides which will either kill them or reduce their longevity thus affecting their 93 vectorial capacity [1]. In sub-Saharan Africa insecticide-treated net (ITN) ownership is estimated 94 to have increased from 3% in 2000 to 83% in the period of 2015-2017 [1]. In Kenya, the 95 government rolled out the universal bed net programme where every two persons in a household 96 were provided with a free ITN. The ITN ownership rose from 12.8% in 2004 to over 80% in Malaria vectors have been shown to adapt to changing environment due to either behavioural 100 avoidance or selection of mutations and recombination that favour their survival in the presence 101 of insecticides threatening the efficacy of the current indoor-based vector control tools [5,11] and 102 the resulting increase in residual transmission [12]. Insecticide resistance is common in sub- 103 Saharan Africa with some regions reporting resistance to all classes of insecticides [13,14]. In 104 Kenya, the target site and metabolic resistance mechanisms play a major role in pyrethroid 105 resistance [15,16]. The primary malaria vectors in Kenya belong to An. gambiae complex and An. 106 funestus group due to their anthropophilic and endophilic behaviours that makes them be more

123
In order to improve vector control intervention strategies, it is crucial to characterize the 124 behavioural patterns of each species of a particular vectorial system in their specific settings over 125 time and in a range of environmental changes, especially with increasing pyrethroid resistance.

126
The objective of this study was to investigate the species diversity of malaria vectors, their resting   Gainesville, FL, USA) from outdoor kitchens and evening outdoor human resting points.

154
Anopheline mosquitoes were sorted morphologically according to the identification keys described 155 by [29]. Female mosquitoes were further classified according to their gonotrophic status.

156
Mosquitoes from each collection method were stored in separately labeled vials and preserved by 157 desiccation.

158
Some of the collected indoor and outdoor resting mosquitoes that were either blood fed or gravid 159 from the two sites were kept in paper cups covered with moistened cotton towels and transported    Anopheles species resting indoors was significantly higher by 82.4% than outdoor location 17.6% 220 across the study sites (z = -8.47, p < 0.0001). The mean indoor resting density of An. gambiae s.l 221 from both sites was significantly higher than outdoor resting density (F 1 , 655 =41.928, p < 0.0001).

269
The Vgsc-1014S was the only kdr mutation observed in An. arabiensis resting indoors (10%) and 270 was not detected in the outdoor resting collections (Table 1).

271
In Kisian, the frequency of Vgsc-1014S in An. gambiae was 70% and that of 1014F was 0.8 %,    In Bungoma, the overall seasonal infective rate was high during the dry season (18%) than in the 319 rainy season (7.6%) for An. funestus. For Anopheles gambiae, the seasonal infective rate was 7.2% 320 during the dry season and 7.06% during the rainy season. Anopheles arabiensis seasonal infective 321 rate was only detected during the rainy season (5.6%) ( Table 3). In Kisian, the seasonal infective 322 rate of An. gambiae was 2.4 % during the dry season.  mosquitoes may be of the particular concern given that the mutation has been found to be strongly 363 associated with pyrethroid resistance than Vgsc-1014S which has been observed to be a weaker  The authors wish to thank the villagers and community leaders in Bungoma and Kisian for their 453 permission to collect mosquitoes in their houses. In particular, we thank the community leaders.   Competing interest 465 The authors declare that they have no competing interest.

466
Availability of data and materials 467 The datasets used for the current study are available from the corresponding author on reasonable 468 request.