We* read with interest this paper by Glunt et al. It contains some interesting ideas on developing a methodology to differentially kill young and old mosquitoes, following the reasoning that killing old mosquitoes, and ideally old malaria- infected mosquitoes, could provide effective malaria control while only weakly selecting for resistance (Read et al, 2009).

However, we would like to quantify and draw attention to what is, in our opinion, the largest danger inherent in this approach, namely increasing the dominance levels of resistant mutations. The dominance of a mutation is not an inherent property, but is determined by the environment within which selection takes place. Denote the resistance mutation as ‘R’ and the original susceptible form as ‘S’. High concentrations may kill both RR and RS genotypes making the ‘R’ mutation recessive, while low concentration may allow the RS genotype to survive, making it dominant (see Figure 1 of Barbosa et al).
This is vitally important because dominance relationships between susceptible and resistance alleles hugely affect the rate of spread of resistance. Increasing dominance greatly increases the rate at which resistance evolves (Figure 2 of Barbosa et al). As a specific example, if initial frequency of resistance is 0.1% and insecticide deployment makes the resistant homozygote 20% fitter than the wild type (s=0.2), we can use the standard population genetic equation (Equation 1 in Barbosa et al) to predict the time for insecticide resistance to spread. If the resistant mutation is near recessive (h=0.1) it takes 248 generations (around 20 years assuming 12 generations per year) to reach an overall frequency of 50%. In contrast, if low insecticide concentrations make it semi-dominant (h=0.5) then it takes 73 generations (around 6 years) and if it is near dominant (h=0.9) it only takes around 47 generations (4 years) to reach 50%. These differences far exceed the differences likely to be generated by their proposal to deploy insecticides at low concentrations. They note these concerns about altered dominance levels stating that it represents “conventional wisdom”. We would prefer the more objective term “elementary genetic theory” which, at least in our opinion, in this case appears to be extremely robust.

Glunt et al discussed a number of practical difficulties in translating their approach into policy, to which we would add the following.
(1)Insecticides applied at low concentrations will decay over time to nearly ineffective levels. The application on surfaces such as walls may also be patchy. This temporal and spatial heterogeneity may well result in mosquitoes being exposed to a mosaic of ineffective, ‘low’ and ‘high’ concentrations and it is not clear how this heterogeneity will affect their conclusions.
(2) It would be difficult to accurately calibrate this strategy because we have no real idea of how exposure in the lab correlates with that in the field. For example Glunt et al used a modified WHO assay where mosquitoes are continuously exposed for 1 hour and it is hard to understand how this will translate into killing in the field where mosquitoes exposure to insecticides on walls may be very prolonged (if they rest on the walls) or extremely brief when mosquitoes may make contact with bednets for only a few seconds.
(3) The experiments were conducted using a single susceptible strain, reared under controlled laboratory conditions, which does not mimic the genetic and environmental variation that occurs in nature. The patterns of survival would most likely change if resistance was already present and the pattern would depend also on the type of resistance (target site or detoxification) (Rajatileka et al, 2010).

In summary, we would note that the paper makes some interesting points but, for the sake of policy makers, would stress that such strategies carry huge dangers in altering the dominance/recessively of insecticide resistance that, at least in our opinion, preclude its practical application in the present form.

* Susana Barbosa and Ian M. Hastings
Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA, England

REFERENCES
Barbosa S, Black WC IV, Hastings I (2011) Challenges in Estimating Insecticide Selection Pressures from Mosquito Field Data. PLoS Negl Trop Dis 5(11): e1387. doi:10.1371/journal.pntd.0001387

Rajatileka S, Burhani J, Ranson H (2011) Mosquito age and susceptibility to insecticides.Transactions of the Royal Society of Tropical Medicine and Hygiene 105(5).

Read AF, Lynch PA, Thomas MB (2009) How to make evolution-proof insecticides for malaria control. PLoS Biol 7(4): e1000058. doi:10.1371/journal. pbio.1000058