| dc.description.abstract | Malaria continues to be a global public health priority with unprecedented scale-up of
insecticide based interventions in the sub-tropics. The World Health Organization (WHO) has
approved vector control using insecticide. treated nets (lTNs) and indoor residual-insecticide
spraying (IRS) as the key malaria vector control strategies and has been adopted in most
countries in Africa. In Kenya, despite the increase in ITN coverage and the expansion of IRS
programs especially in western Kenya, mosquitoes continue to feed a~d, transmit malaria. One
question from this observation is whether the presence of these interventions have influenced
the host preferences for the local vectors or whether some species, due to their feeding habit
and feeding behavior, continue to proliferate and have gained dominance in this region. The
purpose of this study was to investigate host preference by three main malaria vectors (An.
gambiae s.s., An. arabiensis and An. Junestus) in the presence of different levels of vector
control interventions in western Kenya and to assess the accuracy of different mosquito
sampling methods aimed at collecting vectors from the immediate presence of the host in
estimating host selection for Anopheles vectors. The study was conducted in 4 districts;
Nyando, Rarieda, Teso and Bungoma selected based on vector composition and -intervention
type. Mosquito samples were collected indoors (by pyrethrum spray collection and CDC light
traps) and outdoors (by clay pots and CDC light traps) during the months of September to
December, 2010. The samples were identified morphologically and by polymerase chain
reaction (PCR), and all blood fed Anopheles samples were analyzed by sequencing followed by
a Basic Local Alignment Search Tool (BLAST) search in the GenBank database to determine
mosquito blood meal host. Results revealed no significant impact of intervention on selection
for human blood meal for IRS (p = 0.09), both IRS and ITN (p = 0.74) and ITN (p = 0.55).
The range of hosts fed on by Anopheles vectors 'in western Kenya included human (82.57%),
bovine (3.63%), goat (0.65%), rats (0.28%), wild birds (0.09%) and frogs (0.09%). An.
gambiae s.s. and An. Junestus displayed an almost exclusive selection for human blood meal
without a significant difference between the species (p = 0.31), while An. arabiensis fed on
both human and cattle but with significantly higher preference for cattle blood meal (p =
0.0002). From the collection methods, pyrethrum spray collection (PSC) results showed 93.2%
human and 6.08% cattle blood meal, light traps both indoor and outdoor had 97.1% human and
2.8% cattle blood meal while pots reported 100% human blood meal. Information generated
from this study is necessary for improvement of implementation of vector control. Despite
these interventions, the feeding behavior as measured by host selection was similar across all
houses, indicating that interventions present in those houses did not affect host selection
behavior. Two of the three malaria vector species (An. gambiae s.s and An. Junestus) displayed
strong anthropophilic trends while a third (An. arabiensisi fed on humans and cattle. Besides,
host selection was observed to be dependent on vector behavior and was unlikely to be
accurately estimated by mosquito sampling techniques aimed at collecting vectors from a
specific environment of the host. The extent of human-vector contact as depicted in this study
predicted continued risk of malaria infection and adequate measures are necessary to establish
vector feeding behavior, prevention and control. Pyrethrum spray collections appeared to be
the best tool in collection of fed and half gravid samples for blood meal analysis while at the
same time provided samples for the widest range of Anopheles hosts. | en_US |
