A genetic model of the effects of insecticide-treated bed nets on the evolution of insecticide-resistance

Abstract

Background and objectives: The evolution of insecticide-resistance in malaria vectors is emerging as a serious challenge for the control of malaria. Modelling the spread of insecticide-resistance is an essential tool to understand the evolutionary pressures and dynamics caused by the application of insecticides.

Methodology: We developed a population-genetic model of the spread of insecticide-resistance in a population of Anopheles vectors in response to insecticides used either as adulticides (focussing on insecticide-treated bed nets (ITNs)) or as larvicides (either for the control of malaria or, as an inadvertent side-product, in agriculture).

Results: We show that indoor use of insecticides leads to considerably less selection pressure than their use as larvicides, supporting the idea that most resistance of malaria vectors is due to the agricultural use of the insecticides that are also used for malaria control. The reasons for the relatively low selection pressure posed by adulticides are (i) that males are not affected by the ITNs and, in particular, (ii) that the insecticides are also repellents, keeping mosquitoes at bay from contacting the insecticide but also driving them to bite either people who do not use the insecticide or alternative hosts.

Conclusion: We conclude by discussing the opposing public health benefits of high repellency at an epidemiological and an evolutionary timescale: whereas repellency is beneficial to delay the evolution of resistance, other models have shown that it decreases the population-level protection of the insecticide.

Keywords: insecticide-resistance; insecticide-treated bed nets; malaria control; repellency.

Figure 1.

Host searching cycle of a mosquito: A mosquito bites an animal with probability $1-Q$, while a proportion Q of the mosquitoes attempts to bite a human host inside houses, of which a proportion $\varphi$ is protected by ITNs. We assume that mosquitoes survive feeding-associated death, same in humans and animals, with a probability σ. If mosquitoes target a protected house, there are three possible outcomes: the mosquito is repelled by the insecticide (or mechanically blocked by the net) with a probability r, or, if not repelled, it feeds and then escapes the risks of insecticide-associated death with probability s

Figure 2.

The combination of coverage by ITNs and by larvicides that enable resistance to be fixed (lines) or eliminated (below lines) for (a) repellency, r ranging from 0 along the thin line to 1 along the thick line with an interval of 0.2 between adjacent lines and for (b) human feeding, Q, ranging from 1 along the thin line to 0 along the thick line. Other parameter values are given in Table 2

Figure 3.

The rate of evolution of resistance against ITNs as a function of their coverage and for several parameter values. The rate of evolution is given as the time (in number of mosquito generations) it takes for a resistance allele to reach a frequency of 50%, starting with a frequency of 1/100000, (a) The role of a larvicide used simultaneously at a coverage ψ ranging from 0 (thin line) to 25% (thick line). (b) The role of the repellency of the ITN, with repellency r ranging from 0 (thin line) to 0.8 (thick line). (c) The role of the likelihood that mosquitoes feed on animals, with indoor human-feeding Q ranging from 1 (thick line) to 0.6 (thin line). Other parameter values are given in Table 2

Figure 4.

The comparison of the rate of evolution in situations where the insecticide is used only on ITNs or only as larvicides. The y-axis shows the ratio of the two times (in number of mosquito generations) it takes for a resistance allele to reach a frequency of 50%, starting with a frequency of 0.00001 in males and females; the x-axis shows the coverage by the insecticide in either situation. The lines show different levels of repellency, r, ranging from 0 (thin line) to 0.6 (thick line). Other parameter values are given in Table 2