Determinants of optimal insecticide resistance management strategies

Highlights

• Application strategies affect the effective life of insecticide products.

• Strategies for applying two insecticides were tested in a simulation model.

• Mixtures at reduced-dose most often resulted in the longest effective life.

• An insect’s reproduction strategy was the key determinant of the optimal strategy.

Abstract

The use of insecticides to control agricultural pests has resulted in resistance developing to most known insecticidal modes of action. Strategies by which resistance can be slowed are necessary to prolong the effectiveness of the remaining modes of action. Here we use a flexible mathematical model of resistance evolution to compare four insecticide application strategies: (i) applying one insecticide until failure, then switching to a second insecticide (sequential application), (ii) mixing two insecticides at their full label doses, (iii) rotating (alternating) two insecticides at full label dose, or (iv) mixing two insecticides at a reduced dose (with each mixture component at half the full label dose). The model represents target-site resistance.

Multiple simulations were run representing different insect life-histories and insecticide characteristics. The analysis shows that none of the strategies examined were optimal for all the simulations. The four strategies: reduced dose mixture, label dose mixture, sequential application and label dose rotation, were optimal in 52%, 22%, 20% and 6% of simulations respectively.

The most important trait determining the optimal strategy in a single simulation was whether or not the insect pest underwent sexual reproduction. For asexual insects, sequential application was most frequently the optimal strategy, while a label-dose mixture was rarely optimal. Conversely, for sexual insects a mixture was nearly always the optimal strategy, with reduced dose mixture being optimal twice as frequently as label dose mixture. When sequential application of insecticides is not an option, reduced dose mixture is most frequently the optimal strategy whatever an insect’s reproduction.

Introduction

The development of resistance against chemical insecticides in insects appears to be inevitable, with most arthropod pest species having evolved resistance to at least one insecticide (Tabashnik et al., 2014). Insecticide application strategies that can reduce the speed with which resistance builds up are therefore necessary. Since it is generally recognised that applying a single insecticidal mode of action repeatedly over successive applications will lead to the rapid build-up of resistance against that mode of action, using multiple insecticidal modes of actions is assumed to reduce the rate at which resistance builds up, keeping the insecticides effective for longer. However, the way in which two or more insecticides should be combined for optimum resistance management is not well understood. In this article we explore how best to combine two different insecticide modes of action to ensure their effectiveness for the longest possible time.

Four management strategies are frequently considered as ways to combine multiple insecticides (Bourguet et al., 2013). These are: mixtures, where insecticides are applied simultaneously; alternation of each insecticide in time, frequently termed a rotation; separation of each insecticide in space, often called a spatial mosaic; and applying a single insecticide until it becomes ineffective before switching to a different insecticidal mode of action, termed sequential use.

A variety of studies have been carried out in the past to try and determine which strategy (or strategies) can best prevent the development of resistance, including field studies (e.g. Parker et al. (2006)), laboratory studies (e.g. McKenzie and Byford, 1993, Prabhaker et al., 1998) as well as mathematical modelling (e.g. Argentine et al., 1994, Curtis, 1985, Stratonovitch et al., 2014). A recent review from the REX Consortium (Bourguet et al., 2013) that aimed to synergise the available experimental and theoretical studies in antibiotic, insecticide, herbicide and fungicide resistance research, concluded that mixtures in which each component was used at its full label dose (the registered maximum dose per application, specified on the label) was the optimal resistance management strategy. Indeed, for insecticides, 14 out of 16 studies showed that mixtures were the most effective resistance management strategy. The authors highlighted “multiple intragenerational killing” as the reason for this, also known as redundant killing.

There are two classical explanations for why mixtures of two insecticides, applied at full dose, should reduce the selection rate for resistance: redundant killing, and lowering the dominance of the resistant alleles (Curtis et al., 1978). With two modes of action being used simultaneously against an insect pest, redundant killing posits that individuals resistant to one mode of action would be killed by the other mode of action in the mixture and vice versa, thus reducing the level of resistance in the population. Reducing the dominance, on the other hand, posits that by using a high dose of insecticide, the heterozygote individuals in a population will be killed at as high a rate as the sensitive individuals, thus making the insect functionally recessive, irrespective of the true dominance of the resistance allele. With only resistant homozygote individuals surviving an application of insecticide, these resistant homozygote individuals will mate with the remaining sensitive individuals, again creating heterozygotes which will, again and recurrently, be killed by the high dose of pesticide applied (Curtis et al., 1978).

Despite this insecticide mixtures are not often used. There are several reasons for this. Firstly, that the use of mixtures of insecticides at their label dose increases the amount of active ingredient used compared to rotations of insecticides at their label dose, which leads to control above and beyond what is necessary, and greater environmental impact. Additionally, where the insecticides are targeting different insect species that can be damaging at different times in the crop life cycle, the application of insecticide mixtures at the wrong time would be wasteful. The Insecticide Resistance Action Committee (IRAC) advises that rotating different modes of actions is usually the best strategy (IRAC, 2012), and that if mixtures are to be used they should be used at their registered rates, be of different modes of action with no cross-resistance, have little resistance to either mode of action, and have similar periods of residual insecticidal activity (IRAC, 2012).

Most modelling studies either model a single insect and test for sensitivity in parameters (e.g. Stratonovitch et al., 2014, Argentine et al., 1994), or start with a default parameterisation and perform a monofactorial parameter search from here (e.g. Curtis (1985)). While this provides insight around the initial parameter values, the parameter space in reality is considerably larger, and interactions between different life-cycle parameters may influence the effectiveness of different management strategies. A more comprehensive overview of the benefit of each management strategy for multiple insect pests requires a model that can test each strategy in a model pest with different life-history structures, genetics of resistance, life-cycle parameter values, and insecticidal traits. Such a model would enable a global sensitivity analysis from which general conclusions can be drawn.

Therefore we present a flexible, deterministic model describing the selection of target-site resistance in an insect pest population, in order to examine what features of an insect-insecticide system lead to different resistance management strategies being optimal. Four resistance management strategies are compared: application of one insecticide until failure, followed by application of the other (sequential application: SA); a mixture of two insecticides at their label dose (a label-dose mixture: LM); rotating two insecticides over a succession of years at their label dose (label-dose rotation: LR); and a mixture of two insecticides at a reduced dose that leads to a similar efficacy as a single insecticide at label dose (reduced-dose mixture: RM).

To gain the greatest insight, we explore this model in two ways. Firstly, a monofactorial search, starting from two initial model parameterisations, in which the insect life-cycle parameters, genetics, reproduction, as well as the effect of the insecticide and the degree of resistance towards the insecticides are all varied. Secondly, a global analysis in which the model structure and parameter values are generated randomly. In all scenarios we are testing whether a SA, LM, LR or RM results in the slowest selection for resistance to the insecticides in each particular realisation. In a farming context, the success of a resistance management strategy can be measured as the ‘effective life’ of a mode of action against a particular pest species. ‘Effective life’ is used here to convert the effect of selection into a practically meaningful output, quantified as the number of years until loss of effective control.

The aim of the paper is to examine whether particular traits, either of an insect’s life cycle or of the insecticide resistance genes, determine which of the four strategies results in the longest effective control of a given insect pest.