'A bite before bed': exposure to malaria vectors outside the times of net use in the highlands of western Kenya

Abstract

Background: The human population in the highlands of Nyanza Province, western Kenya, is subject to sporadic epidemics of Plasmodium falciparum. Indoor residual spraying (IRS) and long-lasting insecticide treated nets (LLINs) are used widely in this area. These interventions are most effective when Anopheles rest and feed indoors and when biting occurs at times when individuals use LLINs. It is therefore important to test the current assumption of vector feeding preferences, and late night feeding times, in order to estimate the extent to which LLINs protect the inhabitants from vector bites.

Methods: Mosquito collections were made for six consecutive nights each month between June 2011 and May 2012. CDC light-traps were set next to occupied LLINs inside and outside randomly selected houses and emptied hourly. The net usage of residents, their hours of house entry and exit and times of sleeping were recorded and the individual hourly exposure to vectors indoors and outdoors was calculated. Using these data, the true protective efficacy of nets (P*), for this population was estimated, and compared between genders, age groups and from month to month.

Results: Primary vector species (Anopheles funestus s.l. and Anopheles arabiensis) were more likely to feed indoors but the secondary vector Anopheles coustani demonstrated exophagic behaviour (p < 0.05). A rise in vector biting activity was recorded at 19:30 outdoors and 18:30 indoors. Individuals using LLINs experienced a moderate reduction in their overall exposure to malaria vectors from 1.3 to 0.47 bites per night. The P* for the population over the study period was calculated as 51% and varied significantly with age and season (p < 0.01).

Conclusions: In the present study, LLINs offered the local population partial protection against malaria vector bites. It is likely that P* would be estimated to be greater if the overall suppression of the local vector population due to widespread community net use could be taken into account. However, the overlap of early biting habit of vectors and human activity in this region indicates that additional methods of vector control are required to limit transmission. Regular surveillance of both vector behaviour and domestic human-behaviour patterns would assist the planning of future control interventions in this region.

Figure 1

Maps of the study site showing the sampling quadrants, and phases of recruitment. a Construction of sampling grid and identification of building structures using aerial maps; b Survey of sampling grid to identify and exclude quadrants without breeding sites or with fewer than four houses; c Randomization of houses within the remaining quadrants and sequential recruitment of four houses per quadrant; d An example of a typical night of sampling, with six quadrants active and six quadrants deactivated.

Figure 2

Mean hourly catch of a Anopheles arabiensis and b Anopheles funestus s.l. caught by CDC light-traps. Traps were emptied hourly between 17:30 and 22:29 each evening and between 05:30 and 06:29 the next morning.

Figure 3

Monthly mean true protective efficacy of nets (P*) against the combined bites of primary malaria vectors. For the purpose of this study, primary malaria vectors are defined as An. nili, An. funestus s.l. and An. gambiae s.l.

Figure 4

Variation in mean true protective efficacy of nets (P*) by age group of participants. Calculations based on a bed net efficacy where nets are estimated to prevent 80% of bites when used correctly.

Figure 5

Combined hourly man biting rate (MBR) for Anopheles arabiensis and Anopheles funestus s.l. Biting activity overlaid on the reported movements of the local human population indoors and outdoors before, during and after sleep (mean hours). Data for outdoor hourly MBRs were not collected between the hours of 22:30–05:29. For diagrammatic proposes, data for indoor MBR estimates between the hours of 22:30–05:29 were divided equally across the 7 h of collection. Data collected between June 2011 and May 2012.