Mosquitoes and bednets: testing the spatial positioning of insecticide on nets and the rationale...

Mosquitoes and bednets: testing the spatial positioning ofinsecticide on nets and the rationale behind combinationinsecticide treatments

R. M. OXBOROUGH*,{, F. W. MOSHA*, J. MATOWO*, R. MNDEME*, E. FESTON*,

J. HEMINGWAY{ and M. ROWLAND{

*Kilimanjaro Christian Medical Centre, P.O. Box 3010, Moshi, Tanzania{London School of Hygiene and Tropical Medicine, Keppel Street, London WC1E 7HT, U.K.{Liverpool School of Tropical Medicine, Pembroke Place, Liverpool L3 5QA, U.K

Received 23 April 2008, Revised 3 July 2008,

Accepted 7 July 2008

The recent development of pyrethroid resistance of operational significance in Anopheles gambiae s.l. is a major

threat to the control of malaria in West Africa. The so-called ‘2-in-1’ bednet, in which the top of the net is treated

with a non-pyrethroid insecticide and the sides with pyrethroid, has been proposed as a way of maintaining efficacy

in the wake of such resistance. A host-seeking female Anopheles mosquito must contact both the top and sides of a

‘2-in-1’ net, however, for such nets to be useful in resistance management.

In the present study, the interaction between mosquitoes and insecticide-treated bednets (ITN) was explored by

restricting the insecticide to particular surfaces of the nets (top only or sides only) and then testing these nets,

untreated nets and nets treated on all their surfaces in experimental huts, under simulated field conditions. Over the

6-week trial, there was no significant difference in An. arabiensis mortality between nets treated with pyrethroid on

the top only (39.2%), sides only (39.6%) and all surfaces (39.7%), thus indicating that a female An. arabiensis

usually contacts both the top and sides of a bednet during its host-seeking behaviour. The data on blood-feeding

indicated, however, that the insecticide used on the sides of the net may be more important in preventing mosquito

biting than that on the top.

These results support the rationale behind the ‘2-in-1’ nets. Such nets may have advantages over the use of nets

treated on all surfaces with a mixture of insecticides that includes a non-pyrethroid component. With the ‘2-in-1’,

the more toxic component can be deployed on the top of the net, away from human contact, while the more

repellent pyrethroid can be restricted to the sides, to prevent blood-feeding.

With the scaling-up of ITN coverage and the need to preserve pyrethroid efficacy, more consideration should be

given to switching from pyrethroid-only nets to ‘combination’ nets that have been treated with a pyrethroid and

another insecticide. Since the mosquitoes that act as malarial vectors may contact all surfaces of a bednet during

their host-seeking, spatial heterogeneity in insecticide levels over the surface of a net may not reduce that net’s

overall efficacy. Nets with a rather uneven distribution of insecticide (such as those that might be produced using

home-treatment insecticide kits) may therefore be no less effective, prior to washing, than nets with a more even

distribution of insecticide (such as long-lasting insecticidal nets produced under factory conditions).

A key target set out in the global strategic plan

of the Roll Back Malaria Partnership is ‘for

80% of people at risk from malaria to be

protected by 2010 through locally appropriate

vector control such as insecticide-treated nets

(ITN) and indoor residual spraying (IRS)’

(WHO, 2005). To achieve this goal, the

insecticides used to treat bednets or house

surfaces must be efficacious in reducing

mosquito feeding on human blood, by

personally protecting those who sleep under

the ITN or by the community-wide mass

killing of mosquitoes. The main biological

Reprint requests to: R. M. Oxborough, London School ofHygiene and Tropical Medicine, Keppel Street, LondonWC1E 7HT, U.K.E-mail: richard.oxborough@lshtm.ac.uk; fax:z44(0)207299 4720.

Annals of Tropical Medicine & Parasitology, Vol. 102, No. 8, 717–727 (2008)

# 2008 The Liverpool School of Tropical Medicine

DOI: 10.1179/136485908X337553

threat to sustaining malaria control through

use of ITN and IRS is the development of

insecticide resistance in the mosquitoes that

act as malarial vectors.

The gene that confers knockdown

resistance to pyrethroids (kdr) is already

widespread in Anopheles gambiae s.s. in

many areas of western Africa (Chandre

et al., 1999; Etang et al., 2006; Pinto et al.,

2006) and has also been reported in parts of

eastern Africa (Vulule et al., 1994).

Although kdr initially appeared to be no

obstacle to malaria control in many areas

(Henry et al., 2005), the recent emergence

and spread of pyrethroid resistance and/or

kdr in the M form of An. gambiae may

severely limit the effectiveness of ITN and

IRS (N’Guessan et al., 2007; Sharp et al.,

2007). In southern Africa, the emergence of

An. funestus with metabolic pyrethroid resis-

tance is believed to be the main reason why

the malaria burden in Kwazulu Natal rose

seven-fold between 1995 and 1999

(Hargreaves et al., 2000).

It has been suggested that, for vector

control, existing organophosphates and car-

bamates might be suitable alternatives to

pyrethroids (Kolaczinski et al., 2000;

Hougard et al., 2003; Asidi et al., 2004,

2005). As mosquitoes find such insecticides

less irritant and excito-repellent than pyre-

throids, they make longer contact with nets

treated with these insecticides than with

pyrethroid-treated nets (Hougard et al.,

2003). Nets treated with organophosphates

or carbamates therefore tend to cause

greater mosquito mortality than pyre-

throid-treated nets but give less personal

protection (Hougard et al., 2003). In theory

at least, the use of nets that have been

treated not only with a pyrethroid but also

with one of these alternative insecticides

should give high levels of mosquito mortal-

ity and personal protection while reducing

the selection pressure for resistance. If such

‘combination’ nets are to be effective in the

management of resistance, however, each

insecticide component has to kill most of

those mosquitoes that are resistant to the

other component (Mani, 1985; Tabashnik,

1990). The only mosquitoes that should

survive exposure to the nets are the very rare

double mutants that carry resistance to both

insecticides. Theoretical models indicate

that, provided a minority of other mosqui-

toes evades contact with either insecticide

and is free to mate with the rare double

mutants, selection of resistance should still

be slow to evolve (Taylor and Georghiou,

1979). In practice, mixtures of insecticides

work in more subtle ways than can be

predicted using deterministic population

genetics. For example, at high coverages,

an excito-repellent component may stimu-

late pick up of the other insecticide and

enhance mortality (Denholm and Rowland,

1992).

Compared with the use of a mixture of

insecticides on the whole net, the treating of

the roof of a bednet with one insecticide and

the sides with another (to give a so-called ‘2-

in-1’ net) has potential benefits. For exam-

ple, deployment of the more toxic compo-

nent on the roof of the net may reduce any

health risks to those who sleep under the

net. It is suggested that the close proximity

of the two insecticides on the net effectively

means that the two act like a mixture, with

similar resistance-management benefits

(Guillet et al., 2001). As the warm air and

carbon dioxide that emanate from the

sleeper move upwards thermally (Guillet et

al., 2001; Mathenge et al., 2004), the

assumption is that host-seeking mosquitoes

usually explore an occupied bednet from the

top downwards. With a net that has a non-

pyrethroid insecticide on its top and a

pyrethroid on its sides, it might therefore

be expected that a host-seeking mosquito

would pick up a lethal dose of the non-

pyrethroid before being driven away from

the sleeper by the excito-repellent pyre-

throid on the sides. For such ‘2-in-1’ nets

to be useful as a tool for resistance manage-

ment, it is important that the host-seeking

mosquito contacts both the top and sides of

the net, so that a mosquito that is resistant

to one component will still contact the other

718 OXBOROUGH ET AL.

component and be killed by it. The aim of

the present study was to determine whether

host-seeking An. arabiensis do contact both

the top and sides of an occupied bednet, by

comparing mortality and blood-feeding in

experimental huts containing occupied bed-

nets that had been treated on the top only,

on the sides only, or on all surfaces with the

same concentration of a pyrethroid insecti-

cide (lambdacyhalothrin).

MATERIALS AND METHODS

Study Area and Insecticide Treatments

Evaluation of the lambdacyhalothrin-treated

nets was carried out under both field

conditions (in experimental huts at

Mabogini field station, in Lower Moshi,

northern Tanzania, in an area of rice

irrigation) and in a laboratory setting (con-

tact bio-assays were conducted at the

Kilimanjaro Christian Medical Centre, in

Moshi, northern Tanzania). The only sig-

nificant human-biting mosquitoes in the

Lower Moshi area are An. arabiensis and

Culex quinquefasciatus (Ijumba et al., 2002).

Test materials were rectangular polyester

bednets. There were four types of net, with

three nets/type. Three nets were left whole

and not impregnated, as controls. Each of

the rest was cut to separate the top from the

four sides and then the top, the four sides or

the whole net was impregnated, with 18 mg

lambdacyhalothrin/m2, before the net was

sewn back together. Care was taken to

ensure there was no gap in the area of

stitching between the top and sides of each

reconstructed net. The top had an area of

2.9 m2 whereas, together, the four sides

covered 17.1 m2. Three replicate nets were

made for each of the three ‘treatments’ and

the untreated control.

Contact Bio-assay

Each of the 12 mosquito nets was subjected

to contact (cone) bio-assays (WHO, 2006)

on two occasions: after the net was sewn

together but immediately before it was used

in the experimental-hut trials and, again,

several weeks after the field trials had

finished. At each time-point, sugar-fed, 2-

to 5-day-old, laboratory-reared An. arabien-

sis (Dondotha) were tested on each net —

three replicates of five mosquitoes each on

the top of each net and three replicates of

five mosquitoes each on two sides of each

net, giving 45 test mosquitoes/net. The

mosquitoes were exposed to the net surface

for 3 min and then transferred to paper cups

for mortality assessment 24 h later.

Experimental-hut Evaluation

The four experimental huts used in the 6-

week field trial were based on the design

described by Smith (1964) and Smith and

Webley (1969), with some slight modifica-

tions (reduction of eave space, addition of

hardboard ceilings lined with hessian cloth,

replacement of supporting pillars with a

concrete floor surrounded by a water filled

moat, and improved screening of the ver-

andah). The total verandah catch was

doubled to adjust for the loss of mosquitoes

through the open verandahs.

During the trial, the nine treated nets

(treated 2–3 days before) and the three

untreated nets were rotated through each of

the four huts. The volunteers who slept

under the bednets were rotated between

huts on successive nights, in order to reduce

potential bias caused by inter-individual

differences in attractiveness to the local

mosquitoes. The direction of the two open

verandahs in each hut was routinely chan-

ged with each treatment rotation, to mini-

mise the potential confounding effect of a

preferential escape route. None of the nets

was holed during the trial.

Each morning during the trial (at

07.00 hours), mosquitoes were collected

from inside the net, the window (exit) traps,

and the ceiling, walls and floors of the

verandahs and room. They were kept for

species identification, determination of

gonotrophic stage, and mortality counts.

MOSQUITOES AND BEDNETS 719

All members of the An. gambiae complex

identified by morphological characteristics

were assumed to be An. arabiensis, based on

the results of previous cytotaxonomic and

PCR-based identifications of the mosqui-

toes in the study area (Ijumba et al., 2002;

Kulkarni et al., 2006). Mosquitoes were

held in paper cups and provided with 10%

glucose solution for 24 h before their

mortality was scored.

All the data collected were double-

entered and analysed to show the effect of

each type of impregnation (top only, sides

only, or all net surfaces) on the exiting ‘rate’

(proportion, of all of the mosquitoes col-

lected, that came from the verandahs and

exit traps), blood-feeding ‘rate’ (proportion,

of all of the mosquitoes collected, that had

blood-fed) and mortality ‘rates’ [propor-

tions, of all the mosquitoes collected, that

were found dead in the morning (immediate

mortality) and after a further 24 h]. Each of

these outcome variables for each type of

impregnation was compared with the corre-

sponding values for the controls, and also

with each of the other types of impregna-

tion, using logistic regression and version

8.0 of the Stata software package

(StataCorp, College Station, TX). A P-

value of ,0.05 was considered indicative

of a statistically significant difference.

RESULTS

Contact Bio-assays

In the contact bio-assays, which were con-

ducted before the hut trials, similarly high

mortalities (.65%) were observed on all the

lambdacyhalothrin-treated samples, and

relatively low mortalities (,12%) were

observed on all the untreated surfaces

(Fig. 1). This indicates that no cross-con-

tamination had occurred during the treat-

ments and the sewing of the net pieces back

together. Contact bio-assays conducted sev-

eral weeks after the conclusion of the hut

trial still showed high mortality for the net

parts that had been treated with insecticide

(ranging from 87% to 100%), confirming

FIG. 1. The results of the cone bio-assays conducted immediately prior to the experimental-hut trial, showing the

mortality obtained with the tops (%) and sides (&) of the treated and untreated nets. The vertical lines indicate

95% confidence intervals.

720 OXBOROUGH ET AL.

insecticide integrity during the trial period.

By this time, the untreated tops of the nets

that had been treated only on their sides

had, unfortunately, become contaminated

with the insecticide, presumably during

storage after the trial (giving mortality up

to 78%).

Experimental-hut Trials

Anopheles arabiensis

NUMBERS CAUGHT/NIGHT. In general, the

number of An. arabiensis caught each night

in each experimental hut did not differ

significantly between the three types of

treatment or between each type of treatment

and the untreated control. Exceptionally, on

the third night of the trial, five times as

many An. arabiensis were caught in the hut

with one of the fully treated nets (155 in

total) as in any of the other huts. Although

this result did not cause undue bias to the

analysis (non-parametric statistics), it did

skew the total for that particular treatment

(Table 1).

EXITING. In huts with the untreated nets, a

total of 90.8% of the An. arabiensis exited to

the verandahs. In the huts with any of the

treated nets, however, the corresponding pro-

portions were significantly higher (Table 1).

BLOOD-FEEDING. Blood-feeding was signifi-

cantly rarer in the huts with any type of

treated net than in the huts with the

untreated nets (Table 1). Although similar

levels of blood-feeding were seen with the

nets that only had their sides treated (16.2%

blood-fed) as with the nets treated on all

surfaces (12.7%), the latter nets gave a

significantly lower level of blood-feeding

than seen with the nets that only had their

tops treated (18.1%).

MORTALITY. The three types of net treat-

ment induced similar levels of An. arabiensis

mortality 24 h post-exposure, and these

levels were all significantly higher than the

corresponding mortalities seen with the

untreated nets (Table 1). There were no

significant differences in mortality between

the three types of treatment. The temporal

trend seen in mortality over the 6 weeks of the

trial was consistent within each treatment and

did not differ between treatments (Fig. 2).

Culex quinquefasciatus

NUMBERS CAUGHT/NIGHT. The mean num-

bers of Cx. quinquefasciatus seen in the hut

collections ranged from 2.6 to 4.6/night.

Significantly fewer mosquitoes were caught

in the huts with the nets that had been

treated only on their sides than in those with

the nets that had been treated on all of their

surfaces (Table 2). As there is no obvious

cause for this difference and there was no

similar trend in the An. arabiensis collections

TABLE 1. The results of trials of pyrethroid (lambdacyhalothrin) treatments on bednets, against Anopheles ara-

biensis in experimental huts*

Untreated net

Part of net treated, at 18 mg/m2:

Top only Sides only All surfaces

Total no. of females caught 422a 497a 551a 769a

Females caught/night 11.7 13.8 15.3 21.4

OUTCOME VARIABLE (%) AND (95% CONFIDENCE INTERVAL)

Exiting 90.8 (87.6–93.2)a 95.8 (93.6–97.2)b 96.4 (94.4–97.6)b 97.4 (96.0–98.3)b

Blood-feeding 24.6 (20.8–29.0)a 18.1 (15.0–21.7)b 16.2 (13.3–19.5)bc 13.1 (10.9–15.7)c

Blood-feeding inhibition – 26.4 34.1 46.7

Mortality 24 h post-exposure

Observed 9.7 (7.2–12.9)a 39.2 (35.0–43.6)b 39.6 (35.6–43.7)b 39.7 (36.3–43.2)b

Corrected for control mortality – 32.7 33.1 33.2

*Within each row, values sharing the same superscript letter do not differ significantly (P.0.05).

MOSQUITOES AND BEDNETS 721

made at the same time, this difference is

considered to be a ‘type-II’ error (showing a

statistical difference when, in truth, there is

none).

EXITING. Exiting of Cx. quinquefasciatus ran-

ged from 81.5% to 89.1%, with no significant

differences between the three types of net

treatment or between each type of treatment

and the untreated control (Table 2).

BLOOD-FEEDING. Relative to the untreated

control, the nets treated only on their tops

produced the smallest reduction in the

blood-feeding of Cx. quinquefasciatus, and

the nets treated only on their sides produced

the greatest reduction (73.4%). There were,

however, no significant differences between

the three types of net treatment (Table 2).

MORTALITY. Although the 24-h mortality of

Cx. quinquefasciatus with the treated nets

ranged from 19.5% to 27.2% (each type of

treated net giving significantly higher mor-

tality than the untreated control), there were

no significant differences in such mortality

FIG. 2. Changes in mortality of Anopheles arabiensis entering experimental huts over the 6-week trial period,

showing the values recorded in the first 2 weeks (%), third and fourth weeks (&) and last 2 weeks (&). The vertical

lines indicate 95% confidence intervals.

TABLE 2. The results of trials of pyrethroid (lambdacyhalothrin) treatments on bednets, against Culex quinquefas-

ciatus in experimental huts*

Untreated net

Part of net treated, at 18 mg/m2:

Top only Sides only All surfaces

Total no. of females caught 119ab 128ab 92a 166b

Females caught/night 3.3 3.6 2.6 4.6

OUTCOME VARIABLE (%) AND (95% CONFIDENCE INTERVAL)

Exiting 81.5 (73.5–87.5)a 89.1 (82.4–93.4)a 88.0 (79.7–93.3)a 88.6 (82.8–92.6)a

Blood-feeding 24.4 (17.5–32.9)a 13.3 (8.4–20.3)b 6.5 (3.0–13.8)b 12.0 (7.9–17.9)b

Blood-feeding inhibition – 45.5 73.4 50.8

Mortality 24 h post-exposure

Observed 6.7 (3.4–12.9)a 19.5 (13.6–27.3)b 27.2 (19.1–37.1)b 27.1 (20.9–34.4)b

Corrected for control mortality – 13.7 22.0 21.9

*Within each row, values sharing the same superscript letter do not differ significantly (P.0.05).

722 OXBOROUGH ET AL.

between the three types of net treatment

(Table 2).

DISCUSSION

In the present hut trials, the nets treated

only on their tops gave similar 24-h mortal-

ity of An. arabiensis as the nets treated only

on their sides, even though a net treated on

all of its surfaces had a six-fold greater area

of treated material than a net treated on the

top only. There appear to be three possible

reasons for this observation.

The first possibility is that each An.

arabiensis in the trials contacted the sides

or the top of the net (but not both the top

and sides), with an equal chance of contact-

ing each. If this were the case, however, An.

arabiensis mortality with the nets treated on

all of their surfaces should have been

approximately equal to the sum of the

mortality on the nets treated on their sides

only (39.6%) and the mortality on the nets

treated on their tops only (39.2%) — that is,

about 80% (whereas the observed value was

39.7%).

A second possibility is that, at some stage

during the 6-week trial, the tops of the nets

that had been treated only on their sides

became contaminated with insecticide from

the sides of the nets, so that, in effect, the

nets resembled the nets treated on all of

their surfaces. Although some support for

this possibility comes from the results of the

post-trial bio-assays (in which it did appear

that the tops of the nets that had been

treated only on their sides had become

contaminated with insecticide), it seems

likely that most, if not all, of the contamina-

tion occurred after the trial, when the nets

were left folded, in storage, for several

weeks. When mortality in the 6-week trial

was broken down into three fortnightly

periods, mortality with the nets treated only

on their sides showed the same temporal

trend as the mortalities recorded on the

other types of treated net (Fig. 2), with no

indication of the gradual contamination of

the tops of the nets during the course of

the trial.

The third and most likely possibility to

explain the similar An. arabiensis mortalities

with each type of treated net is that each

host-seeking An. arabiensis in the trial

persistently attempted to penetrate the bed-

net to reach the sleeper and, in doing so,

searched over a large area of the net,

including both the top and the sides. Other

studies delving into the workings of ‘2-in-1’

nets failed to demonstrate that each host-

seeking mosquito generally contacts all

surfaces of a net. The earlier studies used

combinations of insecticides that differed in

toxicity, behavioural effects, position on the

net or surface area covered, however, and

such complexity makes it difficult to provide

adequate controls or allow inferences about

mosquito behaviour on and around the net

to be made (Guillet et al., 2001; Hougard et

al., 2003; Asidi et al., 2005).

Although the bednets used in this study

were un-holed, many of the An. arabiensis

caught in the huts containing untreated nets

had blood-fed (presumably on the volun-

teers under the nets). Encouragingly, how-

ever, treatment of the sides and/or top of a

net with lambdacyhalothrin resulted in

significantly fewer blood-fed An. arabiensis,

confirming the importance of pyrethroids

for personal protection (Miller et al., 1991;

D’Alessandro et al, 1995; Asidi et al., 2005).

Interestingly, treatment of all the surfaces of

a net produced significantly fewer blood-fed

mosquitoes than the treatment of the top

only, lending support to the notion that the

insecticide used on the sides of a net is more

important for personal protection than that

used on the top, because the sleeper is more

likely to be in contact with the sides than the

top (the insecticide chosen for treating the

sides should therefore be repellent). In the

present study, the nets treated only on their

sides should have produced proportionately

fewer blood-fed An. arabiensis than the nets

treated only on their tops; although such a

difference was observed, it did not reach

statistical significance.

MOSQUITOES AND BEDNETS 723

If a difference in contact-repellency was

the key factor in reducing An. arabiensis

blood-feeding with the nets treated on all of

their surfaces, compared with that seen with

the nets treated only on their tops, the nets

treated on all surfaces should have been

associated with a greater degree of exiting.

Although such a difference was observed

(Table 1), it also did not reach statistical

significance. Exiting with the untreated

control was already high, however, since

An. arabiensis is exophilic compared with

An. gambiae s.s. (Mahande et al., 2007).

With Cx. quinquefasciatus, treatment only of

the net top produced the lowest mortality of

all treatments (Table 2). This stands in

contrast with the results for An. arabiensis

and may indicate behavioural differences

between the two species.

The present results support the concept of

the ‘2-in-1’ bednet. To achieve resistance

management, the targeted mosquito must

contact both the treated top of a ‘2-in-1’ net

and the treated sides. Wild, host-seeking

An. arabiensis appear to satisfy this criterion

(present study) and therefore, against this

species at least, ‘2-in-1’ treatment should

have a similar impact, in decreasing the risk

of resistance development, to a mixture of

insecticides applied to all surfaces of a net.

This raises the question of whether a ‘2-in-

1’ bednet has any advantages over a net

treated with a mixture. Published studies on

‘2-in-1’ nets have specifically focused on

organophosphates and carbamates, which

are potent inhibitors of human cholines-

terases (Miller et al., 1991; Kolaczinski et al.,

2000; Asidi et al., 2004). The 2-in-1 method

might be a way of reducing health risks to the

sleeper by deploying such non-pyrethroids

as far from the sleeper as possible (i.e. on the

top of the net).

If two insecticides — one that is repellent

and one that is good at killing mosquitoes —

are to be used on a net, it is more appropriate

to use the non-repellent insecticide on the top

of the net, at a dose sufficient to kill the target

insect, and the repellent insecticide on the

sides, to reduce blood-feeding. Resistance

management with mixtures or ‘2-in-1’ nets

works on the principle of redundant killing:

those insects resistant to one component of

the combination will come into contact and

be killed by the other component (Denholm

and Rowland, 1992). Three assumptions

must be met for optimal use of insecticide

combinations (Tabashnik, 1990): (1) the

insect should not be resistant to both

components; (2) the combination must

maintain its integrity over time, with the

components showing similar decay rates; and

(3) the modes of resistance must be unique.

Several insecticides new to public health,

such as chlorfenapyr, have shown potential

in initial trials on nets (N’Guessan et al.,

2007; Mosha et al., 2008). Some older

organophosphates that combine low mam-

malian toxicity and low levels of resistance

to insensitive acetylcholinesterase mechan-

isms also show potential (Hemingway et al.,

1984; Kolaczinski et al., 2000). No alter-

native insecticide has the pyrethroids’ twin

attributes of generating excito-repellency

and high mortality in mosquitoes at low

concentration, however, and hence it is

essential that the pyrethroids be preserved

from the threat of resistance if at all possible.

The combining on nets of a non-pyrethroid

with a pyrethroid would have advantages in

all areas of Africa. In areas where most

mosquitoes are susceptibile to pyrethroids,

the non-pyrethroid component (either in

mixture or on the top of the net) would be

expected to kill any pyrethroid-resistant

mosquito that comes into contact with it,

thereby reducing the selection of pyrethroid

resistance, whereas the pyrethroid compo-

nent should continue to kill susceptible

mosquitoes and provide personal protec-

tion against them. In areas where resistance

is already at high frequency, the non-

pyrethroid component would be expected

to kill resistant mosquitoes and, at high

levels of ITN coverage, to reduce malaria

transmission.

In the present study, insecticide combina-

tions were not explored per se. Rather, the

intention was to show, by effects on

724 OXBOROUGH ET AL.

mortality, how host-seeking mosquitoes

contact bednets and pick up insecticide.

This was achieved through tests involving a

single insecticide restricted to given sur-

faces. To have tested a combination at this

stage would have confused the picture (since

no other class of insecticide induces beha-

viour or toxicity in the same way as the

pyrethroids) and would have made inter-

pretation of the data much more difficult.

Other researchers, such as Asidi et al.

(2005), have gone straight to the combined

testing of two insecticides and this has

tended to cloud the picture rather than shed

light on how each component works. By

using just a single insecticide in the present

study, it was possible to show that host-

seeking females of one member of the An.

gambiae complex tend to roam over all

sections of a bednet, including the top

where alternative insecticides might be put.

This sets the scene for further work on

combinations.

With the scaling-up of ITN coverage under

the Global Fund (www.theglobalfund.org)

and the President’s Malaria Initiative

(www.fightingmalaria.gov), there is a grave

risk of accelerating the selection of pyrethroid

resistance. Consideration should be given to

switching from nets treated with a single

insecticide to nets treated with at least two

insecticides, either in the form of a mixture or

as ‘2-in-1’, to help preserve the essential

resource represented by the pyrethroids.

The present data have other important

implications. Aside from ‘2-in-1’ nets and

the problem of resistance, there is concern

that heterogeneity in the pyrethroid content

on the surfaces of individual nets may

reduce effectiveness. The insecticide on nets

treated at home, by dipping, is more uneven

than seen in factory-produced ITN (Yates et

al., 2005; Hill et al., 2006). Even in the era

of long-lasting insecticidal nets, there

remains a significant market for long-lasting

treatment kits in which the insecticide

formulation is mixed in aqueous solution

with a polymer binder that, once dried on

the nets, protects the insecticide from

removal during subsequent washing

(WHO, 2007). The treatments investigated

here indicate that heterogeneity in the

insecticide level over a net may not impact

upon the mortality generated by that net if,

as seen with An. arabiensis, the target

mosquitoes contact multiple surfaces of the

net while host-seeking. Uniform insecticide

impregnation, although desirable from the

perspective of improving the quality of long-

lasting insecticidal nets, may therefore not

be essential for effectiveness.

ACKNOWLEDGEMENTS. The authors thank C.

Masenga, A. Mtui, E. Philip, E. Tillya,

J. Puya, H. Temba, and R. John, for their

field work, and A. Sanga and R. Athuman,

for insectary support. This project was

funded by the Innovative Vector Control

Consortium.

REFERENCES

Asidi, A. N., N’Guessan, R., Hutchinson, R. A.,

Traore-Lamizana, M., Carnevale, P. & Curtis, C.

F. (2004). Experimental hut comparison of nets

treated with carbamate or pyrethroid insecticides,

washed or unwashed, against pyrethroid-resistant

mosquitoes. Medical and Veterinary Entomology, 18,

134–140.

Asidi, A. N., N’Guessan, R., Koffi, A. A., Curtis, C. F.,

Hougard, J. M., Chandre, F., Corbel, V., Darriet, F.,

Zaim, M. & Rowland, M. W. (2005). Experimental

hut evaluation of bed nets treated with an organopho-

sphate (chlorpyrifos methyl) or a pyrethroid (lamb-

dacyhalothrin) alone and in combination against

insecticide resistant Anopheles gambiae and Culex

quinquefasciatus mosquitoes. Malaria Journal, 4, 25.

Chandre, F., Darrier, F., Manga, L., Akogbeto, M.,

Faye, O., Mouchet, J. & Guillet, P. (1999). Status of

pyrethroid resistance in Anopheles gambiae sensu lato.

Bulletin of the World Health Organization, 77, 230–235.

D’Alessandro, U., Olaleye, B. O., McGuire, W.,

Thomson, M. C., Langerock, R., Bennett, S. &

Greenwood, B. M. (1995). A comparison of the

efficacy of insecticide-treated and untreated bed nets

in preventing malaria in Gambian children.

Transactions of the Royal Society of Tropical Medicine

and Hygiene, 47, 305–309.

Denholm, I. & Rowland, M. W. (1992). Tactics for

managing pesticide resistance in arthropods: theory

and practice. Annual Review of Entomology, 37, 91–112.

MOSQUITOES AND BEDNETS 725

Etang, J., Fondjo, E., Chandre, F., Morlais, I.,

Brengues, C., Simard, F., Nwane, P., Chouaibou,

M. & Ndjemai, H. (2006). First report of knockdown

mutations in the malaria vector Anopheles gambiae

from Cameroon. American Journal of Tropical

Medicine and Hygiene, 74, 795–797.

Guillet, P., N’Guessan, R., Darriet, F., Traore-

Lamizana, M., Chandre, F. & Carnevale. P.

(2001). Combined pyrethroid and carbamate ‘two-

in-one’ treated mosquito nets: field efficacy against

pyrethroid-resistant Anopheles gambiae and Culex

quinquefasciatus. Medical and Veterinary Entomology,

15, 105–112.

Hargreaves, K., Koekemoer, L. L., Brooke, B. D.,

Hunt, R. H., Mthembu, J. & Coetzee, M. (2000).

Anopheles funestus resistant to pyrethroid insecticides

in South Africa. Medical and Veterinary Entomology,

14, 181–189.

Hemingway, J., Rowland, M. W. & Kissoon, K.

(1984). Efficacy of pirimiphos-methyl as a larvicide

or adulticide against insecticide resistant and sus-

ceptible mosquitoes. Journal of Economic Entomology,

77, 868–871.

Henry, M. C., Assi, S. B., Rogier, C., Dossou-Yovo, J.,

Chandre, F., Guillet, P. & Carnevale, P. (2005).

Protective efficacy of lambdacyhalothrin treated nets

in Anopheles gambiae pyrethroid resistance areas of

Cote d’Ivoire. American Journal of Tropical Medicine

and Hygiene, 73, 859–864.

Hill, J., Lines, J. & Rowland, M. (2006). Insecticide

treated nets. Advances in Parasitology, 61, 77–128.

Hougard, J. M., Corbel, V., N’Guessan, R., Darriet,

F., Chandre, F., Akogbeto, M., Baldet, T., Guillet,

P., Carnevale, P. & Traore-Lamizana, M. (2003).

Efficacy of mosquito nets treated with insecticide

mixtures or mosaics against insecticide resistant

Anopheles gambiae and Culex quinquefasciatus in

Cote d’Ivoire. Bulletin of Entomological Research, 93,

491–498.

Ijumba, J. N., Mosha, F. W. & Lindsay, S. W. (2002).

Malaria transmission risk variations derived from

different agricultural practices in an irrigated area of

northern Tanzania. Medical and Veterinary

Entomology, 16, 28–38.

Kolaczinski, J. H., Fanello, C., Herve, J. P., Conway,

D. J., Carnevale, P. & Curtis, C. F. (2000).

Experimental and molecular genetic analysis of the

impact of pyrethroid and non-pyrethroid insecticide

impregnated bednets for mosquito control in an area

of pyrethroid resistance. Bulletin of Entomological

Research, 90, 125–132.

Kulkarni, M. A., Rowland, M., Alifrangis, M., Mosha,

F. W., Drakeley, C., Matowo, J., Malima, R., Peter,

J., Kweke, E., Lyimo, I., Magesa, S., Salanti, A. &

Rau, E. M. (2006). Occurrence of the leucine-to-

phenylalanine knockdown resistance (kdr) mutation

in Anopheles arabiensis populations in Tanzania,

detected by a simplified high-throughput SSOP–

ELISA method. Malaria Journal, 5, 56.

Mahande, A., Mosha, F. W., Mahande. J. & Kweka, E.

(2007). Feeding and resting behaviour of malaria

vector, Anopheles arabiensis with reference to zoopro-

phylaxis. Malaria Journal, 6, 100.

Mani, G. S. (1985). Evolution of resistance in the

presence of two insecticides. Genetics, 109, 761–783.

Mathenge, E. M., Omweri, G. O., Irungu, L. W.,

Ndegwa, P. N., Walczak, E., Smith, T. A., Killeen,

G. F. & Knols, B. G. J. (2004). Comparative field

evaluation of the Mbita trap, the Centers for Disease

Control light trap, and the human landing catch for

sampling of malaria vectors in western Kenya. American

Journal of Tropical Medicine and Hygiene, 70, 33–37.

Miller, J. E., Lindsay, S. W. & Armstrong, J. R. M.

(1991). Experimental hut trials of bednets impreg-

nated with synthetic pyrethroid or organophosphate

insecticide for mosquito control in The Gambia.

Medical and Veterinary Entomology, 5, 465–476.

Mosha, F. W., Lyimo, I. N., Oxborough, R. M., Malima,

R., Tenu, F., Matowo, J., Feston, E., Mndeme, R.,

Magesa, S. M. & Rowland, M. (2008). Experimental

hut evaluation of the pyrrole insecticide chlorfenapyr

on bed nets for the control of Anopheles arabiensis and

Culex quinquefasciatus. Tropical Medicine and

International Health, 13, 644–652.

N’Guessan, R., Corbel, V., Akogbeto, M. & Rowland,

M. (2007). Reduced efficacy of insecticide-treated

nets and indoor residual spraying for malaria control

in pyrethroid resistance area, Benin. Emerging

Infectious Diseases, 13, 199–206.

Pinto, J., Lynd, A., Elissa, N., Donnelly, M. J., Costa,

C., Gentile, G., Caccone, A. & do Rosario, V. E.

(2006). Co-occurrence of East and West African kdr

mutations suggests high levels of resistance to

pyrethroid insecticides in Anopheles gambiae from

Librevillle, Gabon. Medical and Veterinary

Entomology, 20, 27–32.

Sharp, B. L., Ridl, F. C., Govender, D., Kuklinski, J. &

Kleinschmidt, I. (2007). Malaria vector control by

indoor residual insecticide spraying on the tropical

island of Bioko, Equatorial Guinea. Malaria Journal,

6, 52.

Smith, A. (1964). A verandah-trap hut for studying the

house-frequenting habits of mosquitos and for assessing

insecticides. I. A description of the verandah-trap hut

and of studies on the egress of Anopheles gambiae Giles

and Mansonia uniformis (Theo.) from an untreated hut.

Bulletin of Entomological Research, 56, 161–167.

Smith, A. & Webley, D. J. (1969). A verandah-trap hut

for studying the house-frequenting habits of mosqui-

toes and for assessing insecticides. II. The effect of

DDT on behaviour and mortality. Bulletin of

Entomological Research, 59, 3–46.

Tabashnik, B. E. (1990). Modelling and evaluation of

resistance management tactics. In Pesticide Resistance

726 OXBOROUGH ET AL.

in Arthropods, eds Roush, R. T. & Tabashnik, B. E.

pp. 153–182. New York, NY: Chapman & Hall.

Taylor, C. E. & Georghiou, G. P. (1979). Suppression

of insecticide resistance by alteration of gene dom-

inance and migration. Journal of Economic

Entomology, 72, 105–109.

Vulule, J. M., Beach, R. F., Atieli, F. K., Mount, D. L.,

Roberts, J. M. & Mwangi, R. W. (1994). Reduced

susceptibility of Anopheles gambiae to permethrin

associated with the use of permethrin-impregnated

bednets and curtains in Kenya. Medical and

Veterinary Entomology, 8, 71–75.

World Health Organization (2005). Roll Back Malaria

Global Strategic Plan 2000–2015. Geneva: Roll Back

Malaria Partnership Secretariat.

World Health Organization (2006). Guidelines for Testing

Mosquito Adulticides for Indoor Residual Spraying and

Treatment of Mosquito Nets. Document WHO/CDS/

NTD/WHOPES/GCDPP/2006.3. Geneva: WHO.

World Health Organization (2007). Report of the Tenth

WHOPES Working Group Meeting, WHO/HQ,

Geneva, 11–14 December 2006. Review of: Spinosad

0.5% GR & 12% SC, Lambda-cyhalothrin 10% CS,

KO-Tab 1-2-3, Interceptor. Document WHO/CDS/

NTD/WHOPES/2007.1. Geneva: WHO

Yates, A., N’Guessan, R., Kaur, H. & Rowland, M.

(2005). Evaluation of KO Tab 1-2-3: a wash-

resistant ‘dip-it-yourself’ insecticide formulation for

long-lasting treatment of mosquito nets. Malaria

Journal, 4, 52.

MOSQUITOES AND BEDNETS 727