Monitoring mosquitoes in urban Dar es Salaam: evaluation of resting boxes, window exit traps, CDC light traps, Ifakara tent traps and human landing catches

Abstract

Background: Ifakara tent traps (ITT) are currently the only sufficiently sensitive, safe, affordable and practical method for routine monitoring host-seeking mosquito densities in Dar es Salaam. However, it is not clear whether ITT catches represent indoors or outdoors biting densities. ITT do not yield samples of resting, fed mosquitoes for blood meal analysis.

Methods: Outdoors mosquito sampling methods, namely human landing catch (HLC), ITT (Design B) and resting boxes (RB) were conducted in parallel with indoors sampling using HLC, Centers for Disease Control and Prevention miniature light traps (LT) and RB as well as window exit traps (WET) in urban Dar es Salaam, rotating them thirteen times through a 3 × 3 Latin Square experimental design replicated in four blocks of three houses. This study was conducted between 6th May and 2rd July 2008, during the main rainy season when mosquito biting densities reach their annual peak.

Results: The mean sensitivities of indoor RB, outdoor RB, WET, LT, ITT (Design B) and HLC placed outdoor relative to HLC placed indoor were 0.01, 0.005, 0.036, 0.052, 0.374, and 1.294 for Anopheles gambiae sensu lato (96% An. gambiae s.s and 4% An. arabiensis), respectively, and 0.017, 0.053, 0.125, 0.423, 0.372 and 1.140 for Culex spp, respectively. The ITT (Design B) catches correlated slightly better to indoor HLC (r(2) = 0.619, P < 0.001, r(2) = 0.231, P = 0.001) than outdoor HLC (r(2) = 0.423, P < 0.001, r(2) = 0.228, P = 0.001) for An. gambiae s.l. and Culex spp respectively but the taxonomic composition of mosquitoes caught by ITT does not match those of the indoor HLC (χ(2) = 607.408, degrees of freedom = 18, P < 0.001). The proportion of An. gambiae caught indoors was unaffected by the use of an LLIN in that house.

Conclusion: The RB, WET and LT are poor methods for surveillance of malaria vector densities in urban Dar es Salaam compared to ITT and HLC but there is still uncertainty over whether the ITT best reflects indoor or outdoor biting densities. The particular LLIN evaluated here failed to significantly reduce house entry by An. gambiae s.l. suggesting a negligible repellence effect.

Figure 1

Window exit trap before fixing to a house window (A), and after installation (B).

Figure 2

Schematic illustration of a typical night's experimental set up (arrangement 1 as illustrated in figure 3) at one location with two blocks, one of which has occupants using untreated nets while the other has participants using long-lasting insecticidal nets (LLINs).

Figure 3

Schematic presentation of three possible arrangements of trapping methods rotated in order through the three houses in any given block. Note that the letters in blue circle represent the identifier within the block for each house, specified as 1A, 1B and 1C for houses in block 1 and 2A, 2B and 2C in for houses in block 2.

Figure 4

Correlation and density dependence of Ifakara tent trap (ITT) [22,23]model B sampling efficiency relative to human landing catch (HLC). The correlation between the catches of An.gambiae s.l. and Culex spp with ITT and HLC plotted in absolute number is presented in left panels with complete equivalence depicted by the diagonal line. Right panels illustrate density dependence as catches in ITT divided by the sum catches of ITT and HLC against the absolute catches of the HLC reference method.