4. '
h

,1

- ‘
Qfin .x
. t x v.-

   

,., .-,<m~vy_» AA‘L

 

'Ir

 

 

24-.

Mr: .
#335?
h

w
.

 

..

*7".

'3. A.....,...
'r

“‘7:

4.x
SI-

.

3‘!-
I
J.

 

THESIS

2:30 ¥

This is to certify that the
dissertation entitled

QUANTITATIVE BEHAVIORAL ANALYSES OF THE
INTERACTIONS OF ANOPHELES GAMBIAE S.S. GILES
(DIPTERA : CULICIDAE) WITH INSECTICIDE-TREATED

BEDNETS

presented by

FRED ANANGWE AMIMO

has been accepted towards fulfillment
of the requirements for the

Ph.D. degree in ENTOMOLOGY

 

 

 

Major'Professor’s Signature

‘flgfl Z3657

Date

/,

MSU is an affirmative-action, equal-opportunity employer

-.-.-‘—.-.-—.—.-.—-.--—-.an.—--.-.--u-----.—.--—.--.---.-.-----.---.-.—.--.-.--.--o---.-.—A—----.

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

DATE DUE DATE DUE DATE DUE
“WWW;

 

 

 

 

m

NUQPI} 23.82%

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:lCIRC/DateDue.indd-p.1

If

QUANTITATIVE BEHAVIORAL ANALYSES OF THE INTERACTIONS OF
ANOPHELES GAMBIAE S.S. GILES (DIPTERA: CULICIDAE) WITH INSECTICIDE-
TREATED BEDNETS

By

Fred Anangwe Amimo

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Entomology

2007

ABSTRACT

QUANTITATIVE BEHAVIORAL ANALYSES OF THE INTERACTIONS OF
ANOPHELES GAMBIAE S.S. GILES (DIPTERA; CULICIDAE) WITH IN SECTICIDE-
TREATED BEDNETS

By

Fred Anangwe Amimo

Direct behavioral observations were conducted on Anopheles gambiae s.s.
females interacting with Permanet-, Olyset- and Netprotect-Treated BedNets (ITNs).
Conditions progressed in complexity from small static cages and a wind tunnel with a
fully-untreated, fully—treated, and half-treated cage and screen nettings, respectively, to a
house-scale four-choice arena in Western Kenya, fitted with treated vs. untreated bednets.
In laboratory tests, each female mosquito was introduced in the arena during the last 2 h
of photophase and observed continuously for 30 min; frequencies, durations, and
locations of behavior were recorded using the Observer sofiware (Noldus, http:
www.noldus). Two hundred 2-3 day old female mosquitoes were released in a central
courtyard of the house arena and their room occupancy, as well as 8 h and 24 h mortality
was recorded. In the small arenas all ITNs elevated flying behavior. Mosquitoes showed
no evidence that they could detect insecticide from a distance or avoid approaching it. In
the wind tunnel they flew upwind with or without insecticide in the plume of host cues.
Mosquitoes occupied rooms equally with or without treated nets. Both 8 h and 24 h
mortality was significant. Overall, this research demonstrated that the mode-of-action of

ITNs is lethality, not avoidance.

ACKNOWLEDGEMENTS

I wish to express my gratitude to the Institute of Intemational Education, working
through the United States of America Embassy, Nairobi and with the Kenyan government
and the University of Eastern Africa, Baraton all through which I was awarded a
Fulbright scholarship to undertake my doctoral studies at Michigan State University. 1
single out Dr. Justus Mbae for his direction and preparation for the program. I am most
grateful and deeply indebted to my co-supervisors, Drs Edward Walker and James R.
Miller for their intellectual stimulation, instruction, and thorough supervision throughout
my dissertation studies. I thank them as Principal Investigators of the project to evaluate
insecticide-treated bednets in an area of Western Kenya sponsored by the National
Institute of Health grant and their granting me the opportunity to be part of this
invaluable project.

I thank my committee members: Drs. Rufus Isaacs, Ke Dong, and Linda
Mansfield for their thorough examination and critical improvements to my dissertation,
as well as their time and participation in all convened meetings. I will ever be grateful for
the teaching from lecturers in the Entomology Department with the able leadership of Dr.
Rich Merritt. I also thank the research and technical staff in this Department for their
assistance. I single out Juan Huang, Piera Siegert, Bill Morgan, Blair Bullard and the
undergraduate students for helping in provision of rearing materials and techniques,
constructive contributions they made on my proposal, and helping in the analysis and

presentation of my data.

iii

I am thankful to Dr. John Vulule, the Director of Center for Vector Biology and
Control Research of Kenya Medical Research Institute in Kisian where I did my actual
research work for the chance and for physical and material provisions in collaboration
with Michigan State University and the Centers for Disease Control. I also thank all staff
and student in the Entomology Section of the Center. I single out George Olang,
Insectary staff Richard Amito and Gideon Nyansikera, and other staff on ground, Charles
Ochieng, William Ochieng among others.

Last but not least I thank my dear wife, Catherine Adhiambo Amimo, who gave
me full support to go for this program and stayed with me through the demanding times
between the USA and Kenya. My children Windy Amimo, Wendy Amimo, and Ricky

Amimo were very inspirational in untold ways and I thank them as much.

iv

TABLE OF CONTENTS

LIST OF TABLES .................................................................................. vii

LIST OF FIGURES ................................................................................ viii

CHAPTER 1: BEDNETS AS A TACTIC FOR MALARIA SUPPRESSION:

LITERATURE OVERVIEW

Impact of Malaria in Africa ................................................................ 1
Mosquito Vectors of Malaria ............................................................... 2
Tactics for Malaria Treatment and Suppression ......................................... 3
Pyrethroid Insecticides Incorporated into ITNs .......................................... 6
Tests of ITN Safety and Efficacy ......................................................... 10
Concerns about Resistance of Mosquitoes to ITNs .................................... 14
Possible ITN Modes-of-Action and Means for Distinguishing among Them. . ....16
ITN Behavioral Effects on Mosquitoes .................................................. 17
Rationale ..................................................................................... 22
Hypotheses .................................................................................. 23
Research Objectives ......................................................................... 23

CHAPTER TWO: QUANTITATIVE ANALYSES OF BEHAVIORAL
INTERACTION OF FEMALE ANOPHELES GAMBIAE S.S.
GILES WITH INSECTICIDE-TREATED NETTING IN SMALL
STATIC CAGES AND A WIND TUNNEL

Introduction .......................................................................................... 24

Materials and Methods ............................................................................. 25
Mosquitoes ................................................................................... 25
Experimental Series 1 — Tests of Insecticide-Treated and Untreated Netting in
Small Static Cages .......................................................................... 27
Experimental Series 2 — Tests of Untreated and Insecticide-Treated Netting in
a Wind Tunnel with Host Cues ............................................................ 29
Wind Tunnel Procedure .................................................................... 31
Data Analysis ................................................................................ 32

Results ................................................................................................. 33
Experimental Series 1 — Tests of Insecticide-Treated and Untreated Netting in

Small Static Cages ......................................................................... 33
Test A — Untreated Cage vs. Fully-treated Permanet, Olyset, and Netprotect
Cages ......................................................................................... 33
Test B — Untreated Cage vs. Half-treated Permanet, Olyset and Netprotect

Cages ......................................................................................... 38
Experimental Series 2 - Tests of Untreated and Insecticide-Treated Netting in

a Wind Tunnel with Host Cues ........................................................... 49
Test A - Untreated cage vs. fully-treated Permanet, and Olyset screens ............ 49

Test B - Untreated cage vs. half-treated Deltamethrin, and Perrnethrin Screens...65

Discussion ............................................................................................ 68
Main Pre-knockdown Effects of ITNs on An. gambiae Behaviors .................. 68
Any Evidence for Avoidance? ............................................................................... 69
Mode of Action of ITNs ................................................................... 70
Fit of Current Data with Previous Reports .............................................. 70
CHAPTER THREE: QUANTIFYING REDISTRIBUTION AND

SURVIVORSHIP OF ANOPHELES GAMBIAE IN A
MULTI-ROOM CHOICE ARENA PRESENTING
HUMAN SLEEPERS UNDER IN SECTICIDE-TREATED

VS. UNTREATED-BEDNETS

Introduction .......................................................................................... 72
Materials and Methods .............................................................................. 75
Apparatus .................................................................................... 75

Test protocol ................................................................................. 79

Room Equivalent Tests and Effect of Opening Height ............................... 80
Experimental Design for Testing ITNs .................................................. 80

Effect of An. gambiae Strain and Age ................................................... 81
Variations in Site of Release .............................................................. 81

Data Analysis ................................................................................ 82
Results ................................................................................................ 82
Temperature/RH Profiles .................................................................. 82

Room Equivalence Tests and Effect of Height Opening ............................. 84

Effects of ITN Type ........................................................................ 85

Effect of An. gambiae Strain and Age ................................................... 90

vi

Variation in Release Site ................................................................... 92

Netprotect .................................................................................... 95
Discussion ............................................................................................ 99
Performance of Test Apparatus .......................................................... 99

Eave Openings Functioned Strongly as Inlets, and Weakly as Outlets. . . . . . . . 100

ITN Mode-of-Action: Avoidance vs. Lethality ....................................... 101

Slight Differences with ITN Brand .................................................... 103
Recommendation ................................................................................... 1 03
REFERENCES ..................................................................................... 104

vii

LIST OF TABLES

Table 2.1. Comparison of elapsed time for first knockdown event for An. gambiae
females in small cages fully- or half-treated with ITN fabrics ...................... 39

Table 2.2. A 10 x 10 transition matrix of behavioral responses of An. gambiae
when untreated netting covered the upwind end of a wind tunnel ................ 58

Table 2.3. A 7 x 7 transition matrix of positions of An. gambiae behaviors when
untreated netting covered the upwind end of a wind tunnel .......................... 59

Table 2.4. A 10 x 10 transition matrix of behavioral responses of An. gambiae
when Permanet covered the upwind end ofa wind tunnel......... .....59

Table 2.5. A 7 x 7 transition matrix of positional placement of An. gambiae when
Perrnanet covered the upwind end of a wind tunnel ................................... 60

Table 2.6. A 10 x 10 transition matrix of behavioral responses of An. gambiae
when Olyset covered the upwind end of a wind tunnel .............................. 60

Table 2.7. A 7 x 7 transition matrix of positions of An. gambiae behaviors when
untreated netting covered the upwind end of a wind tunnel .......................... 61

vii

LIST OF FIGURES

Figure 1.1 Global distribution of malaria risk. Note: Sub-Saharan Africa experiences
the greatest risk of malaria epidemics, as well as the highest death rates ................... 1

Figure 1.2. Interior of a house typical of the Luo tribe near Kisumu, Kenya. The ca. 30-
cm high opening at the top of the walls and under the roof over-hang is continuous
around the circumference of these houses. It offers mosquitoes unencumbered entrance
into and exit from houses ............................................................................. 3

Figure 1.3. Insecticide-treated-bednets (ITNs) hanging over two different beds in a
Western Kenyan home. A priority of ITN implementations is malaria-protection for
young children and pregnant mothers; malaria can be transmitted in utero .................. 6

Figure 1.4. Chemical structures of six pyrethroid insecticides .................................. 7

Figure 1.5. Location of a long-running CDC and Michigan State University ITN study
site relative to the location of the city of Kisumu. Kenya. This map shows the Center

for Vector Biology and Control Research / Kenya Medical Research Institute in

Kisian where most of the research reported in this dissertation was conducted. . . . .. 1 2

Figure 1.6. Expansion of Dethier et al.’s (1960) terminology for effects of chemicals
on the locomotor behaviors of insects, now structured to show the relationships
among various types of effectors .................................................................. 18

Figure 2.1. Representation of a 20 cm diam. x 20 cm tall static cage used for
observations of An. gambiae females responding to ITNs. Cages were either

constructed: 1) entirely of untreated polyester netting, 2) of half-untreated netting

and the other half of Permanet, or 3) half-untreated netting and the other half Olyset.
halved cages were split vertically ................................................................. 28

Figure 2.2. Wind tunnel for testing responses of An. gambiae females to insecticide-
treated bednet materials placed downwind of host cues ....................................... 29

viii

Figure 2.3. Average percent time An. gambiae allocated to various behaviors during
the 30-min observations of single females in untreated static cages or cages fully-
treated with deltamethrin (Netprotect and Permanet) and permethrin (Olyset) ............ 34

Figure 2.4. Total time spent in given locations in cage during 30-min observations
of An. gambiae females in static cages treated with three types of ITN fabric ............. 35

Figure 2.5. Mean total frequencies and mean total durations for behavioral responses

of An. gambiae females broken out per quarter of the 30 min observation period.

This test was conducted using either fully-untreated cage walls and roof or fully-

treated cage walls and roof ..................................................................... 36

Figure 2.6. Mean total frequency of arrival at particular positions and mean total
durations of stay at those locations for female An. gambiae in static cages, as

broken out by quarter of the 30 min observation period. This test was conducted

using either fully-untreated cage walls and roof or fully-treated cage walls and roof ..... 37

Figure 2.7. Mean total frequencies for behavioral responses of An. gambiae females
broken out per quarter of the 30 min observation period. This test was conducted
using either fully-untreated cage walls and roof or fully-treated cage walls and roof.. . ..40

Figure 2.8. Mean total frequencies for knockdown of An. gambiae females broken
out per quarter of the 30 min observation period. This test was conducted using
either fully-untreated cage walls and roof or fully-treated cage walls and roof. . . . . . ....41

Figure 2.9. Average time An. gambiae allocated to various behaviors during the 30-
min observations of single females in untreated static cages or cages half-treated
with deltamethrin (N etprotect and Permanet) and permethrin (Olyset) ..................... 42

Figure 2.10. Average times An. gambiae females spent in particular locations in
untreated vs half-treated static cages as averaged over the whole 30 min observation
period ................................................................................................. 43

Figure 2.11. Mean total frequencies and mean total durations for behavioral

responses of An. gambiae females broken out per quarter of the 30 min observation
period. This test was conducted using either fully-untreated cage walls and roof or
half-treated cage walls and roof .............................................................. 44

ix

Figure 2.12. Mean total frequencies of arrival per side by An. gambiae females

broken out per quarter of the 30 min observation period. This test was conducted

using either fully-untreated cage walls and roof or half-treated cage walls and

roof .................................................................................................... 45

Figure 2.13. Mean total duration of stay by An. gambiae females on treated vs.
untreated sides of half-treated cages broken out per quarter of the 30 min
observation period ................................................................................... 47

Figure 2.14. Mean total frequency and duration of behaviors by An. gambiae
females on treated vs. untreated sides of half-treated cages broken out per quarter of the

30
min observation period .............................................................................. 48

Figure 2.15. Mean total frequency of knockdowns by An. gambiae females on
treated vs. untreated sides of half-treated cages broken out per quarter of the 30 min
observation period ................................................................................... 49

Figure 2.16. Average time An. gambiae allocated to various behaviors during the 30-
min observations of single females in a wind tunnel with untreated and fully ITN-
treated upwind screens, as averaged over the full 30-min observational period. . .. . . ...50

Figure 2.17. Average times An. gambiae spent in particular locations during the 30-
min observations of single females in a wind tunnel with untreated and fully ITN-
treated upwind screens, as averaged over the full 30-min observational period. . . . . . .51

Figure 2.18. Mean total frequencies and mean total durations for behavioral

responses of An. gambiae females broken out per quarter of the 30 min observation
period. This test was conducted in a wind tunnel using either fully untreated or fully
ITN-treated upwind screens. Data within a column sharing the same lower-case letter

are not statistically significant; data across quarters sharing the same capital letter

are not statistically different ........................................................................ 53

Figure 2.19. Mean total frequencies and mean total durations for behavioral

responses of An. gambiae females broken out per quarter of the 30 min observation
period. This test was conducted in a wind tunnel using either fully untreated or fully
ITN-treated upwind screens ........................................................................ 54

Figure 2.20. Mean total frequencies of anival by An. gambiae females on the

upwind screen of a wind tunnel, broken out per quarter of the 30 min observation

period. This test was conducted in a wind tunnel using either fully untreated or fully
ITN—treated upwind screens ........................................................................ 56

Figure 2.21. Mean duration of visits by An. gambiae females to the upwind screen

of a wind ttmnel, broken out per quarter of the 30 min observation period. This test

was conducted in a wind tunnel using either fully untreated or fully ITN-treated

upwind screens. Data within a column sharing the same lower-case letter are not
statistically significant .............................................................................. 56

Figure 2.22. Mean total frequencies of arrival by An. gambiae females on the
down-wind screen of a wind tunnel, broken out per quarter of the 30 min observation
period. This test was conducted in a wind tunnel using either fully untreated or fully
ITN-treated upwind screens ........................................................................ 57

Figure 2.23. Kinematic graphs of behavioral sequences of An. gambiae in a wind
tunnel with upwind screens configured immediately downwind of host cues .............. 64

Figure 2.24. Mean total frequency of arrivals by An. gambiae on upwind screen in a
wind tunnel having screens either fully untreated or half-treated with deltamethrin or
permethrin, as broken out by quarters of the 30-min observational period .................. 66

Figure 2.25. Mean total durations of visits by An. gambiae on upwind screen in a
wind tunnel having screens either fully untreated or half-treated with deltamethrin or
permethrin, as broken out by quarters of the 30-min observational period .................. 67

Figure 3.1. Photograph of the Mosquito Hotel on site, showing the main entrance
to the central courtyard, and the exterior of two of four sleeping rooms ..................... 75

Figure 3.2. Floor plan of the Mosquito Hotel showing the configuration of the
central courtyard and sleeping rooms ............................................................ 76

Figure 3.3.a. Simulated eave opening above the entrance of a sleeping room .............. 78

xi

Figure 3.3.b. Entrance restrictor with screening and movable bar for adjustment of
cave opening to 3, 10 or 20cm. The smaller of the two openings projected into the
sleeping room ......................................................................................... 78

Figure 3.4. Plots of Temperature and RH in the sleeping rooms on a single test day. . ....83

Figure 3.5. Plots of Temperature and RH of sleeping rooms on different dates and
test months of 2006 .................................................................................. 84

Figure 3.6. Mean percent An. gambiae occupancy in rooms with untreated nets for
simulated eave opening height of 20 cm. Two hundred mosquitoes were released on
each of 12 nights of testing ........................................................................ 85

Figure 3.7.a. Mean percent An. gambiae occupancy in rooms with untreated nets for
simulated eave opening heights of 3, 10 and 20 cm. Two hundred female mosquitoes
were released in the central courtyard each of 12 nights of testing ........................... 86

Figure 3.7.b. Mean percent An. gambiae mortality in rooms with untreated nets for
eave opening heights of 3, 10 and 20cm ..................................................................... 86

Figure 3.8.a. Mean occupancy of female An. gambiae in sleeping rooms with
Permanet and Olyset vs. untreated Bednets. Two hundred mosquitoes were released
in the central courtyard on each of 14 nights of testing ......................................... 87

Figure 3.8.b. Eighth mortality (A) and 24 h mortality (B) of female An. gambiae in
rooms with Permanet, Olyset and untreated nets ................................................ 87

Figure 3.9. Mean occupancy of female An. gambiae in rooms with Permanet, Olyset
and F endona vs. untreated bednets. Two hundred mosquitoes were released in the
central courtyard on each of 12 nights of testing ................................................ 88

Figure 3.10. Eighth mean percent mortality (A) and 24 h mortality (B) of female An.
gambiae in rooms with Permanet, Olyset, and F endona ITNs vs. an untreated bednet. . .89

Figure 3.11. Mean occupancy of female An. gambiae in rooms with 30 mg
deltamethrin and 55mg chessboard deltamethrin vs. untreated bednets. Two hundred
mosquitoes were released in the courtyard on each of 16 nights of testing .................. 89

xii

Figure 3.12. Mean 8 h percent mortality (A) and 24 h mortality (B) of female An.
gambiae in rooms with 30 mg deltamethrin and 55mg chessboard deltamethrin
ITNs ................................................................................................... 90

Figure 3.13. Mean occupancy and mortality of the Kisumu laboratory strain vs. F1
progeny of wild-caught An. gambiae females in rooms with untreated bednets ........... 91

Figure 3.14. Mean percent occupancy for 4 day vs. 16 day-old female An. gambiae.
two hundred mosquitoes were released in the courtyard on each of 12 nights of
testing ................................................................................................. 91

Figure 3.15. Mean percent occupancy of female An. gambiae in rooms with
Permanet vs. untreated nets (A) and with Olyset vs. untreated nets (B) when 50
mosquitoes were released into each room on each of 8 test nights ........................... 92

Figure 3.16. Mean percent mortality of female An. gambiae in rooms with treated vs.
untreated bednets. A and B = 8 h mortality; C and D = 24 h mortality, respectively. . ...93

Figure 3.17. Occupancy and mortality of An. gambiae females in rooms with eaves
open or closed ...................................................................................... 94

Figure 3.18. Distribution of An. gambiae females in rooms with 20 cm eave openings
when all 200 mosquitoes were released in only 1 of 4 sleeping rooms ..................... 95

Figure 3.19. Occupancy and mortality of An. gambiae females in rooms treated with
Permanet and Netprotect vs. untreated nets. A = 8 h mortality, B = 24 h mortality.

Two hundred mosquitoes were released in the courtyard on each of 16 nights of

testing ................................................................................................. 96

Figure 3.20. Occupancy and mortality of Anopheles females in rooms treated with
Netprotect vs. untreated nets ...................................................................... 97

xiii

Figure 3.21. Occupancy and mortality of An. gambiae females in rooms treated with
Netprotect vs. untreated nets before and after washing. A = 8 h mortality, B = 24 h
mortality. Two hundred mosquitoes were released in the central courtyard on each

of 16 nights of testing .............................................................................. 98

xiv

CHAPTER 1

Bednets as a Tactic for Malaria Suppression: Literature Overview

Impact of Malaria in Africa

Malaria, HIV AIDS, tuberculosis, and childhood diarrhea are the four most
important infectious diseases of humans worldwide (USAID 1997; Molyneux 2004).
Over 70% of the world's 3 billion people live in the ca. 100 countries where malaria is
endemic (Figure 1.1), and 18% live in epidemic-prone zones. An estimated 300-500
million clinical cases of malaria are reported annually; 90% of them occur in Sub-
Saharan Africa (WHO 2000). Here, case fatality rates can be as high as 40%, and the

armual continental death rate is typically 1.5 to 2.7 million.

 

   
 
    

3'; no malarla
I malaria rlsk

 

Figure 1.1. Global distribution of malaria risk. Note: Sub-Saharan Africa
experiences the greatest risk of malaria epidemics, as well as the

highest death rates. Source of Map: WHO.

Beyond human deaths, illness due to malaria impedes economic development.
Per-annum estimated costs for Sub-Saharan Africa alone range from 83-12 billion and
represent a drag on economic grth by as much as 1.3% each year (Gallup and Sachs
2001). Sachs and Mallaney (2002) observe that malaria prospers most were human
societies prosper least, i.e., malaria is a correlate of poverty.

Mosquito Vectors of Malaria

All four known human malaria parasites, Plasmodium falciparum, P. vivax, P.
ovale and P. malariae, are transmitted between humans by female Anopheles mosquitoes
(WHO 1983). Newly emerging adult mosquitoes are always parasite-free; they acquire
malaria parasites only after feeding on the blood of infected humans. In high-burden
areas of Africa, P. falciparum, the most dangerous species of human malaria, causes the
majority of infections and the majority of deaths. The mosquitoes, Anopheles gambiae
Giles and An. fimestus, are the most efficient and prevalent vectors of malaria throughout
Sub-Saharan Africa. An. gambiae is the greatest threat because its populations can
quickly build during the rainy season to resting densities as high as several
hundreds/house/night (J oshi et al., 1975; Lindsay et al., 1998). Although only about 3%
of female An. gambiae in Western Africa have infective oocysts at a given time (Fabian
et al., 2005), the high biting rate raises the probability of malaria infection to near
certainty after just a few days of normal human exposure during the rainy season.
Mosquitoes can easily enter and exit houses because screening of house eve-openings
(Figure 1.2) necessary for heat dissipation via ventilation is unaffordable. Moreover,
female anophiline mosquitoes can live for six weeks or more (Oketch et al., 2003). It is

well known that vector age is highly and positively correlated with vectorial capacity

(Bockarie et al., 1995). In epidemic-prone zones, the combination of a dangerous
pathogen efficiently vectored by high and widely distributed populations of long-lived
mosquitoes results in nearly ubiquitous exposure of humans to malaria parasites.
Tragically, ca. 1 million African children, mostly under 4 years of age, succumb to

malaria each year (WHO, 1995).

",4 “W .1" ‘ "1 zit S- 0‘4. _‘ 3'
_ .. -' ”saw:- “ w / ‘
u “—d r!

 

Figure 1.2. Interior of a house typical of the Luo tribe near Kisumu, Kenya.
The ca. 30-cm high opening at the top of the walls and under the roof over-
hang is continuous around the circumference of these houses. It offers

mosquitoes unencumbered entrance into and exit from houses.

Tactics for Malaria Treatment and Suppression
The most common form of malaria remediation has been the administration of
anti-malaria drugs to sick patients. Beginning with quinine, a long series of different

pharmaceuticals has been employed as malaria treatments and prophylactics. Drug

shifting is necessitated by development of resistance by the parasite in given geographical
regions (Caraballo and Rodriguez-Acosta, 1999). Currently the drugs sulfadoxine-
pyrimethamine and artemisinin and its derivatives are widely used in Africa. The most
prevalent prophylactics taken by visitors to high-risk areas are malarone and doxycycline;
both are intended for temporary usage.

Widespread and on- going administration of anti-malaria drugs to malaria-endemic
regions could theoretically reduce malaria incidence dramatically. However, this practice
has not been adopted because: a) there are concerns that all effective drugs could be lost
due to wide-spread selection for resistance, b) pharmaceutical companies have not shown
an interest in manufacturing the necessary quantities of drugs, probably for monetary
reasons, and c) the world has not rallied the political and economic will to commit to
such an large-scale and continuing humanitarian endeavor.

Considerable research has been devoted to development of vaccines against
malaria (Webster and Hill, 2003). Although various formulations have reached large-
scale field tests, none has proven practical to date. The weak link has been that immunity
lasts for a maximum of several months. Thus, the human population targeted for
protection must be vaccinated repeatedly. This task has been deemed unrealistic due to
the immense cost. Nevertheless, the search for better vaccines continues.

Encouragingly, anti-vector tactics can reduce human suffering and deaths from
malaria. In principle, such tools should be inexpensive and implemented in a sustainable
strategy with a public health rather than medical-treatment mentality. The first such
successful program in Africa commenced in 19405 and it involved regular sprays of

interior house-walls with DDT (Musawenkosi et al., 2004). This program ran for over 10

years and reduced malaria incidence and human suffering appreciably. Moreover, the
environmental damage associated with this particular DDT application was low because
of containment within the dwellings. However, this program was terminated before it
reached its lofiy goal of global malaria eradication (Sachs and Malaney, 2002). Key
reasons were: a) stigma associated with use of DDT and other persistent insecticides, and
b) the socio-economic retrenchment of Great Britain and other world leaders during the
19705. Interestingly, house-sprays with DDT are being revisited (Curtis, 2002), this time
driven by African rather than foreign governments.

Use of bednets for physical protection of humans from mosquito bites has
developed over the past three decades as the most-favored anti-vector tactic for
developing countries. The practice of fully surrounding the beds of sleepers with netting
so as to preclude visits by insects has been known since the 6th Century BC (Lindsay and
Gibson, 1988). However, the perceived potential for this approach rose dramatically in
the 19305 with the realization that bednet fabrics could be safely treated with highly
effective chemical insecticides. If Insecticide-Treated-bedNets (ITNs) (Figure 1.3) were
to kill most mosquitoes alighting upon them, the benefits of such a lethal trap might
extend to other sleepers in the house not under a net, or even to unprotected persons in
nearby houses. This would reduce the density and mean age of local populations of
mosquito vectors (Magesa et al., 1991), thus reducing vectorial capacity (Gary and
Foster, 2001). ITN programs in epidemic-prone zones have become the center-piece of
current malaria-suppression efforts, e.g., the global partnership under the Roll Back
Malaria (RBM) program (Nabarro and Taylor, 1998). Results of such programs will be

summarized following an introduction to ITN technologies.

 

Figure 1.3. lnsecticide-treated-bednets (lTNs) hanging over two different
beds in a Western Kenyan home. A priority of ITN implementations is
malaria-protection for young children and pregnant mothers; malaria can be

transmitted in utero. The source of picture was CDC.

Pyrethroid Insecticides Incorporated Into I INS

Pyrethroids are a group of stable synthetic insecticidal compounds, equivalent of
the natural insecticide, pyrethrum, which are structurally derived from pyrethrins,
naturally found in the flowers of Chrysanthemum cinerariaefolium. The first pyrethroid
insecticide, allethrin, was synthesized in 1948 following the structural elucidation of six
major insecticidal constituents in 19205 and 19305 (Casida, 1980) (Figure 1.4). Currently

there are over 42 such compounds of commercial interest; they are either stereoisomers

or differ totally in their chemical structures (NPTN, 1998). While these insecticides have
low toxicity to mammals, they are lethal to insects at low doses, and, toxicity increases as
temperature decreases i.e. they have a negative temperature coefficient for toxicity
(Chang and Plapp, 1983). Their mode of action is by blocking sodium channels so as to
produce repetitive firing of neurons. Cross-resistance between pyrethroids and DDT has

been demonstrated using some insects that were resistant to both types of chemical.

Hindi SEQ-“Ute

Allethrin Permethrin
N 0
°' °' PM
“WE . © © WW 3' Q
Cypermethrin Deltamethrin
cr
0 “'0 G
0 CF
firm a}
Fenvalerate FMC 51785

Figure 1.4. Chemical structures of six pyrethroid insecticides.

This has been attributed to insensitivity of the nervous system, a mechanism which is
referred to as kdr (knockdown resistance) (Scott and Matsumura, 1981; Dong et al.,

1998). Based on chemical structure, pyrethroids have been divided into Types I and II.

Type I pyrethroids (for example, permethrin, tefluthrin, and allethrin) do not contain an
alpha-cyano group in their molecule. Their action is characterized by restlessness,
incoordination, and inactivity in insects, and these insecticides cause tremors in mammals
(T-syndrome). Type II pyrethroids (for example, deltamethrin, cypennethrin, and
fenvalerate) contain an alpha-cyano group and produce ataxia, convulsions,
hyporesponsiveness in insects and choreathetosis and salivation in mammals (CS-
syndrome) depending on the animal model used (Aldridge, 1990; Tordoir et al., 1994;
Soderlund et al., 2002). However, some toxicologists would rather not consider any
given pyrethroid as Type I or Type II, but look at them as existing along a continuum,
with certain ones eliciting primarily a Type I response (e. g. allethrin, permethrin) and
others like deltamethrin, eliciting a Type II response.

Pyrethroids have been extensively employed for agricultural pest control.
Recently, they have also played a major role in vector-control programs where over 520
tons of active ingredients are utilized annually on a global scale (Zaim & J arnbulingam,
2004). This growing usage in public health is attributed to their high insecticidal potency
at low dosages, rapid knockdown effects, and their relative safety for humans. Typical
LD50 for vertebrates are 250-4000 mg/kg body weight. Unlike many other insecticides,
pyrethroids do not significantly accumulate in the human body; 90 % of administered
dosage is usually excreted in urine and faeces in the first week after exposure (Aldridge,
1990; Vijverberg & van den Bercken, 1990; IPCS, 2000b). However, when pyrethroids
were introduced in the early 19805 in China, cotton growers who handled these

compounds without taking precautions experienced deltamethrin and fenvalerate

poisoning (WHO, 1990). Since 1988 no clinical cases of occupational pyrethroid
poisoning have been reported.

ITNs are impregnated with the pyrethroid insecticides deemed suitable under the
recommendation of the WHO Pesticide Evaluation Scheme (WHOPES). So far, six
pyrethroids have been so recommended: alpha-cypermethrin, cyfluthrin, deltamethrin and
lambdacyhalothrin(a1pha-cyano pyrethroids), and etofenprox and permethrin (non-cyano
pyrethroids) (Hougard et al., 2003).

Initially, bednets were treated by the users, who steeped untreated nets in a
solution of local water in which was dissolved tablets of an emulsified formulation of
pyrethroid provided with an untreated bednet. Such formulations lasted only about 6
months and required regular re—treatment (fi'equently not done as prescribed).
Irnpregnating the netting with insecticides during manufacturing was a major advance in
this technology. The major steps for incorporating pyrethroids into netting in factory
manufacture include: running suitable netting through an aqueous bath containing an
emulsified formulation of pyrethroid, removing excess liquid by squeezing the fabric
between rollers at a prescribed pressure, and quickly drying the treated fabric at a
prescribed temperature. The treated netting is then cut into pieces and sewn into
rectangular or tent-like ITNs of various sizes, the most common of which covers what
westerners refer to as a "single bed." The current, most widely marketed factory-
manufactured ITNs are OlysetTM (l g permethrin/m2 polyethylene fiber; Sumitomo, Inc.)
and PermanetTM (55 mg wash-proof deltamethrin/m2 polyester fabric; Vestegaard
Frandsen, Inc.) (WHO, 2000; N’Guessan et al., 2001). These bednet brands are termed

"long-lasting" because they are reputed to kill mosquitoes at sustained levels for more

than 3 years. However, even these ITNs eventually need to be re—treated or replaced, as
any net becomes worn and tattered with constant use.
Tests of I TN Safety and Efficacy

ITNs were tested as a malaria-intervention tool throughout the 19805. Initially,
their safety for humans was assessed by longitudinal trials overseen by WHO. From the
mid-19805, clinical trials were carried out to assess ITN impact on malaria incidence. In a
late 19805 study in a Gambian village where all nets were treated with permethrin, a
remarkable 72% reduction in malaria parasitaemia was realized. Based on this
encouraging outcome, field tests were extended to four large WHO/T DR trials (Lengeler,
2000) in Burkina F aso, Ghana, Kenya, and The Gambia, again with positive outcomes.
The results led to the Abuja declaration at the Afiica Summit on Roll Back Malaria
(RBMCP 2000) calling for high levels (> 60%) of bednet coverage in malaria-prone
regions. This goal was soon achieved in selected sites, commencing with a notable social-
marketing program in an area of high malaria transmission in Tanzania (Schellenberg et
aL,2001)

Further documentation of ITN efficacy was forthcoming. A series of randomized,
controlled trials in four large areas of stable malaria transmission in The Gambia, Ghana,
Burkina-Faso, and Kenya in the 1990's demonstrated a 20 - 25% reduction in infant
deaths and improvements in the overall health of children (D’Alessandro et al., 1995;
Binka et al., 1996; Nevill et al., 1996; Habluetzel et al., 1997). An average of 6 lives per
1000 children was reported to be saved in the age group 1-59 months each year (Lengeler
2000). Considering that there are about 80 million children under the age of five in

Africa, over 400,000 deaths could be prevented annually. In a further trial conducted in

10

Coastal Kenya in the early 19905, heavy ITN implementation halved the number of
episodes of clinical malaria and cases of severe malaria arriving at the coastal hospitals
(Nevill et al., 1996).

A notable ITN study anchored by the USA Centers for Disease Control has
focused on long-term ITN effectiveness and impact at the community level. This
Western Kenyan study site is located near the city of Kisumu (Figure 1.5), a region
notorious for holoendemic malaria. Approximately 120,000 people have been under
permethrin-impregnated bed nets in a 500-km2 area centered on Asembo and Gem. Key
outcomes relating directly to malaria were that transmission was reduced by over 90% in
intervention villages (Gimnig et al., 2003a). Cross-sectional and longitudinal studies of
children enrolled in the study demonstrated that children had a 19%, 39% and 44%
reduced incidence of parasitemia, anemia, and clinical malaria, respectively (ter Kuile et
al., 2003 a). Furthermore, infants residing in intervention villages experienced better
height and weight gain (mean difference in Z-score = 0.36) compared to those residing in
control villages (ter Kuile et al. 2003b). All-cause mortality in children less than 5 years
was reduced by 20%. This effect was strongest in children less than 1 year with an
estimated 34 lives saved per 1000 infants protected by permethrin-treated bed nets
(Phillips-Howard et al., 2003b).

Entomological evaluation during this experiment was paired with laboratory and
behavioral studies to assess differences in the effects of bednets on different mosquito
species and to assess the importance of prompt re-treatment and actual use of bednets
each night (Gimnig et al., 2003a). The proportion of An. gambiae 5.5. (feeds mainly on

humans) relative to An. arabiensz's (feeds mainly on cattle) was higher in ITN - villages

ll

 

   
   

Siaya
CVBCRlKEMRI at

Kisian Equator
I ’7 - Provincial
4- Hospital

5

 

 

ombasa

 

 

Figure 1.5. Location of a long-running CDC and Michigan State University ITN
study site relative to the location of the city of Kisumu. Kenya. This map shows
the Center for Vector Biology and Control Research / Kenya Medical Research
Institute in Kisian where most of the research reported in this dissertation was
conducted. This map was obtained courtesy of the USA. Centers for Disease
control.

(55 %) than control villages (45 %). This shifi is most parsimoniously attributed to a

steeper reduction in An. gambiae than An. arabiensis populations in the ITN-zone. An

 

 
 
  
 

 

CVECRIKEMRI at
Kisran Equator

I

 

. .. Provincial
1- Hospital

" Kisumu

  
  
 

 

ombasa

 

Figure 1.5. Location of a long-running CDC and Michigan State University lTN
study site relative to the location of the city of Kisumu. Kenya. This map shows
the Center for Vector Biology and Control Research / Kenya Medical Research
Institute in Kisian where most of the research reported in this dissertation was
conducted. This map was obtained courtesy of the USA Centers for Disease
control.

(55 %) than control villages (45 %). This shift is most parsimoniously attributed to a

steeper reduction in An. gambiae than An. arabiensis populations in the ITN-zone. An

funestus populations were suppressed even more than An. gambiae populations because
the former species was much more susceptible to permethrin than the latter. For An.
gambiae, bednets were most effective when re-treated every 6 months and when used
every night. For An. funestus, indoor resting densities were consistently 85-90% lower in
intervention houses compared to control houses, regardless of whether the bednets had
been retreated within the previous 6 months or properly deployed the night before
(Gimnig et al., 2003a).

Measuring mosquito spatial distributions over time relative to ITN- vs. non-ITN
houses in the study yielded clues to ITN mode-of-action when combined with use of
Colombian curtains that estimate mosquito entrance and exit from houses. An. gambiae,
An. arabiensis and An. funestus were all less likely to obtain a blood meal and more
likely to exit from a house with a bednet. Additionally, An. funestus was less likely to
enter a house with a bednet (Mathenge et al., 2001). All-night biting collections
conducted in intervention and control villages yielded little evidence for a shift in biting
times of either An. gambiae or An. funestus (Mathenge et al., 2001).

In addition to demonstrating significant personal protection afforded to bednet
users, spatial analyses of the CDC data supported a community-wide benefit from
bednets. Fifty eight per cent fewer anophelines were collected in houses within 300 m of
an intervention relative to a control village. The suggested interpretation was that
presence of bednets suppressed mosquito populations sufficiently to reduce biting
pressure on occupants of nearby unprotected houses, and that this phenomenon could
occur over appreciable distances (Gimni g et al., 2003b). Spatial analysis of child

morbidity and mortality provided further support of this hypothesis. A strong protective

l3

effect (Relative change in Odds ratio of 7.1,7.8,11.7,20.0 and 4%) was detected in control
houses lacking bednets but located within 300 meters of intervention villages for child
mortality, clinical malaria, moderate anemia, hi gh—density parasitemia, and hemoglobin
levels respectively (Hawley et al., 2003a). This “community effect” on nearby houses
without nets was nearly as strong as the effect observed within villages with bednets.
Together, these results indicate that, for optimal efficacy, insecticide-treated bednets must
cover a high proportion of households, be retreated at adequate intervals, and be deployed
every night (Hawley et al., 2003b).

Collectively, the case is compelling for widespread implementation of high-
density ITN use throughout Sub-Saharan Afiica and beyond. ITNs, as they exist today,
have the capability for saving millions of human lives. Moreover, this life-saving
potential of ITNs should not be squandered.

Concerns about Resistance of Mosquitoes to I T Ns

The hope of any pest management program is that the tools effecting pest
suppression will remain effective over time — i.e. sustainability. History has repeatedly
taught entomologists that continual use of a single control tactic (e.g. a given insecticide
or class of insecticides) is likely to soon lead to resistance in the pest population, be it
metabolic, target site modification, or behavioral. Thus, despite the confidence with which
ITN programs are being implemented in developing countries, forethought and planning
need to be given with respect to how to manage ITN use to maximize sustainability.

Resistance to pyrethroid insecticides is already well-known in insects. It can occur
via mutations in the sodium channel, the target site of these insecticides (knockdown

resistance or kdr) (Dong et al., 1998; Liu et al., 2002), and increased rates of insecticide

l4

detoxification due to up-regulation of certain detoxifying enzymes such as cytochrome
P4505 (Hemingway and Ranson, 2000). Alternatively, the pyrethroids incorporated into
ITNs might cause mosquitoes to begin avoiding treated surfaces or treated houses,
perhaps at the expense of unprotected sleepers. It is not out of the question that blood-
feeding by mosquitoes capable of detecting and avoiding ITNs could shift from night-
time to the evening and morning, when occupants of houses are not under ITNs.
However, one study (Mathenge et al., 2001) failed to detect this effect. Several authors
have already sounded cautions about evolution of resistance of Anopheles mosquitoes to
pyrethroids (Snow et al., 1997; Smith et al., 2001). Lines (1996) argues that resistance
studies should be a research priority, whether field-resistance is high or not. Moreover,
foci of pyrethroid resistance are known to occur in An. gambiae, in both East and West _
Afiica (Vulule et al., 1996; Chandre et al., 1999), and in An. funestus in South Africa
(Hargreaves et al., 2000). Data from experimental huts documented that pyrethroid-
resistant mosquitoes from West Africa were less irritated by pyrethroids, and hence
tended to acquire a higher dose of insecticide than the susceptible insects (Chandre et al.,
2000). Somewhat counter-intuitively, mortality was higher in the resistant population
under exposure to pyrethroid-treated surfaces. Scientists at WHOPES (WHO Pesticide
Evaluation Scheme) have recognized the potential for evolution of mosquito resistance to
ITNs and have offered bi-treated bednets (two different types of insecticides on differing
parts of one bednet) and alternative pyrethroids such as bifenthrin as solutions to
resistance development (Guillet et al., 2001; Hougard et al., 2003). However, before such
measures are taken to side-step development of metabolic and target site resistance, it

seems important that the possible and actual modes of action of lTNs be delineated.

15

Possible I T N Modes-of-A ction and Means for Distinguishing among Them

Surprisingly little is settled about the mechanisms whereby ITNs actually protect
htunan sleepers. Few researchers have focused on the detailed behaviors of mosquitoes as
they interface with ITN surfaces. The possibility initially favored by researchers in this
field was that ITNs enveloping human sleepers act as lethal traps. The chemical (C02 and
body odors) and physical (heat) cues emanating from houses and humans were
envisioned to attract hungry mosquitoes into houses and onto the treated netting adjacent
to a sleeper. After spending sufficient time on treated netting, it was assumed that enough
toxicant would be absorbed for knock-down (falling over due to inability to maintain
normal body stance) and kill soon thereafter. Some mosquitoes might recover from
knockdown, but their lifespan might be subsequently shortened by the stress of the
encounter with toxicant. Indeed, death and shortening of life-span is evident when
mosquitoes are confined within small plastic cones and forced to sit mainly on
insecticide-treated netting (WHO, 1983). However, evidence for ITN lethality to
mosquitoes in houses under normal circumstances is sparse to non-existent, probably
because of: a) the difficulty in visually detecting dead mosquitoes on the normally
cluttered house floors, and b) presence of scavengers like ants that can remove dead
mosquitoes.

An alternative class of explanations for ITN mode—of-action is that presence of
insecticide-contaminated surfaces alters mosquito behavior so that lethality is avoided.
Furthermore, sleepers in a house might experience some behaviorally-mediated
protection, even if they are not under a bednet. Because the literature in medical

entomology is fraught with ambiguities and miss-uses of the terminologies elsewhere

16

codified by formal behaviorists, we offer a new set of behavioral terms and definitions
for use in medical entomology (Figure 1.6). These terms and definitions will be used
consistently throughout this dissertation. Insofar as possible, the non-standard and
sometimes confusing terminologies often used by mosquito researchers will be translated
into the terms of Figure 1.6.

As specified in Figure 1.6, there are diverse mechanistic possibilities whereby
behavioral effectors like the insecticides in ITNs might influence mosquito locomotory
behaviors. They range from simply increasing the frequency of un-directed behaviors
normally expressed in the non-stimulated state (locomotor stimulant), to highly directed
behaviors, e. g., consistently navigating away from the source of stimulation (repellent).
The terms and definitions in the top half of Figure 1.6 are consistent with precedents set
by Dethier (1960) and Kennedy (1977). Overall outcomes are indicated in the bottom half
of our scheme. The novel terms avoidant and engagent are proposed for effectors that
decrease and increase interactions of responders with the source of stimulation,
respectively. Use of these inclusive, results-based terms is recommended when the end-
result of stimulation is known, while the actual mechanisms mediating that outcome are
unknown. Conversely, we recommend that the mechanistically specific terms be used
only when evidence is obtained that they apply.

I TN Behavioral Effects on Mosquitoes

Although not yet offering mechanistic explanations as precise as those invited by

Figure 1.6, there is mounting evidence that ITNs elicit the spectrum of effects represented

in the left half of this diagram. For example, Lines et al., (1987) and Snow et al., (1987)

17

Figure 1.6. Expansion of Dethier et al.’s (1960) terminology for effects of
chemicals on the locomotor behaviors of insects, now structured to show the
relationships among various types of effectors. The upper part of this figure
having Arial font and solid lines refers to movement mechanisms (causes)
and their measurable characteristics. The lower part of this figure having
dotted lines and Comic Sans font indicates behavioral outcomes (effects). In
the interest of simplicity, This scheme does not cover all theoretically possible
mechanisms and outcomes. Movers are expected to shift mechanisms
through time, as well as in response to varying stimulus sets. Active space
refers to the area or volume where the concentration of stimulus is detectable

by the mover.

18

«02332 S
3.3a». 5.5 :thuuku‘RL EARLS—SQ

 

9...qu nexus S . "2:2 .0le “week“.

 

3 :uEuhnMLLWV
3.5%. 5.3. :anuokuskL .389.th

 

006nm u>3uo E u .uEI. uLoE “venom

 

 

fl Eozoaom _ _ «:mLooLu/LL

 

 

 

 

9:3. :Lsou
Lo EoimLmEsoo

8.3-9. .5
E0693:

 

6:8

_ uCNawwtx‘ _

 

unfliEzm can! 2:3 233
EOE 5.} LOB—La .3003 .630.
«a 33.5 .onavoLL on! nova—amfi

tetuuLutLL 5 033.? «us o2

Lent.» «St S 093th Lu: 0?

 

“UK: 0.0 {k t!

Funk... give 5

 

Emuoxw # _ E9550

 

 

Eomnw
womoocuou m. 33E: cog? «an.

E: .0 :OaaLau 2E? 0. 9:32 noszEé
mom-29... coaflaEan .0 0:3 co=n_:E=n Lo ocou 2

E2256

 

5560004

300% .5sz .a 3:25
Sconce: 20E mooaama

 

 

 

8:3
mick flooEv.
3523: E v.55 225.89% 53928 55> A _ v mmcimo: E mmochucE 55>
new
eon—QED

 

3 Loucoamoh
9:250 2:355

 

 

19

suggest that ITNs function by driving visiting mosquitoes out of houses via physical
irritation. By our interpretation, such irritation could increase house-exiting simply by
increasing non-directed locomotion (excitant or deterrent). By analogy, increasing the
temperature of a gas in a leaky container will increase gas exit, albeit by a non-directed
mechanism. Molecules of the gas continue to move randomly, but, they will do so with
greater velocity. Hence, their probability of hitting a gap in the container increases
relative to slower-moving molecules. Such irritative or excitatory effects of pyrethroids,
along with responses causing dispersal are widely recognized in the literature (Davidson
1953; Mouchet & Cavalie, 1961; Lockwood et al., 1984; Threlkeld, 1985; Roberts and
Andre 1994; Chareonviriyaphap et al., 1997; Rutledge etal., 1999; Chareonviriyaphap et
al., 2001; Roberts et al., 2002), Attempts at measuring these effects are appropriately
included in standard bioassays of mosquito responses to ITNs (Coluzzi 1963; Rachou et
a1. 1963; Shalaby 1966, WHO 1970; Bondareva et a1. 1986; Pell et al., 1989; Quinones
and Suarez 1989; Ree and Loong 1989; WHO 1998). More precise research is needed to
distinguish whether these effects represent only locomotor stimulants, or also deterrents
(reduce the rate of return of responders to the zone of stimulation) or true repellents. We
also need to know whether such effects involve only contact stimuli, and/or odors
(volatiles).

Recent field research using arrays of experimental huts housing sleepers under
ITNs has shifted prevailing notions of ITN mode of action from the lethal-trap hypothesis
to avoidance. For example, Miller et al., (1991) reported that ITNs impregnated with
permethrin reduced house-entry by 60%. Likewise, Mathenge et al., (2001) found that

permethrin-impregnated ITNs reduced or prevented house-entry of An. funestus, and

20

increased the exit rates of An. gambiae and An. funestus. No difference was apparent in
the rates of blood feeding and exiting of An. arabiensis. A number of other studies offer
similar findings and interpretations (Lindsay et al., 1991; Takken 2002, Gimnig et al.,
2003). Some authors interpret these effects as being odor-mediated and repellency.
However, evidence for such interpretations is arguable. Lindsay et al., (1991) caution that
such avoidance may be caused by the solvent vapors or some other ingredients associated
with the pyrethroids used in some IT Ns rather than with the given pyrethroids
themselves. Because of their high molecular weights (see Figure 1.4), pyrethroids tend to
have very low vapor pressures. Pauluhn (1999) have noted that some formulations of
pyrethroid insecticides have airborne effects due to movement of dust particles bearing
toxicant. The speculation can be added that pyrethroid molecules, once evaporating, may
rapidly condense upon dust particles and other solid surfaces. Gut et al., (2004) have
documented such adsorptive behaviors by insect pheromone molecules of high molecular
weight.

The oldest studies of mosquito behavioral responses to insecticides are still
informative and worth considering. They were conducted using DDT. Anopheline and
culicine mosquitoes resting on surfaces sprayed with DDT flew away after 5-10 minutes
(WHO, 1963). Mean elapsed time to first departure decreased as DDT concentrations on
paper increased (Brown, 1958; Ungureanu & Teodorescu, 1963). In a study by Shalaby
(1965) excitability/deterrency of a DDT-resistant strain of An. culicifacies Giles did not
differ from that for a DDT-susceptible strain. In contrast, what was reported as DDT
deterrency by an Indian An. stephensi strain resistant to DDT was less than that for a

susceptible strain (Choudhury & Rahman, 1967). Current research (Roberts and Alecrim,

21

1991; Chareonviriyaphap et al., 1997; Roberts et al., 2000) is re-exploring the
pronounced ability of DDT to cause mosquitoes to avoid DDT-treated surfaces; however,
little progress has been made in explaining these effects in terms of Figure 1.6.

A major conclusion of this literature review is that, despite the importance of
ITNs and the considerable research attention they have drawn to date, sizable gaps
remain in our knowledge of how they actually function and whether formulations should
maximize avoidance vs. lethality or vice versa, or be some compromise between the two.
With respect to prevention of transmission of non-zoonotic diseases like malaria,
diversion of mosquitoes from humans to alternate hosts may be a desirable outcome of
any intervention. Interestingly, Bogh et al., (1998) reported that, following introduction
of permethrin-impregnated bed nets in villages in Kwale District, near the coast of
Kenya, the human blood index (proportion of blood meals taken from humans vs. other
animals) dropped from 99.5% to 87.5% in An. gambiae 5.5. sample specimens. This is a
remarkable result, since An. gambiae 5.5. is a highly anthropophilic mosquito reported to
rarely feed on hosts other than humans (Takken & Knols 1999). On the other hand, direct
kill of mosquitoes vectoring malaria directly so as to reduce density and mean vector age
may be a more robust outcome because this reduces vectorial capacity. Research helping
to differentiate among these possibilities should ultimately prove useful to improving and
sustaining malaria suppression via ITNs.
Rationale

ITNs are currently making a major contribution to malaria prevention. Knowledge
of the mode-of—action of IT N5 will be critical to predicting possible response of An.

gambiae populations in response to heavy ITN implementation.

22

Hypotheses

This study postulates that quantitative analyses of malaria mosquitoes interacting
with ITNs will reveal behaviors not entirely explained by toxicosis. This study also
postulates that attraction to host cues and repellence from ITNs will have offsetting
effects on mosquito foraging for a blood meal resulting in continued interaction and

leading to toxicosis.

Research Objectives

The core objectives of this dissertation research are to: characterize, quantify, and
classify the behaviors of blood-seeking An. gambiae interacting with untreated vs. long—
lasting pyrethroid-treated -bednetting currently being deployed in Sub-Saharan Africa.
This research aims for: a) a level of behavioral resolution surpassing that of previous
studies, and b) differentiation among the behavioral responses outlined in the top half of
Figure 1.6. The insights resulting from this study are expected to significantly increase
understanding of the mode-of-action of ITNs. In turn, such knowledge will hopefully
enable leaders of long-term programs for malaria vector suppression to make informed

decisions about how ITNs might be managed to maximize sustainability.

23

CHAPTER 2
Quantitative Analyses of Behavioral Interaction of Female Anopheles gambiae 5.5.

Giles with Insecticide-Treated Netting in Small Static Cages and a Wind-tunnel

INTRODUCTION

Insecticide—treated-bednets (ITNs) with formulations of deltamethrin
(Permanet)TM or permethrin (OlysetTM ) lasting up to five years currently serve as a key
component of the malaria-suppression programs being implemented across Sub-Saharan
Afiica, where malaria has been killing nearly 1 million people annually (WHO, 2000). It
has been proposed that ITNs protect sleepers by several of the following mechanisms: a)
restraining mosquitoes from approaching sufficiently close to human skin to take a blood-
meal, b) killing mosquitoes sitting on the nets sufficiently long to pick up a lethal dosage
(Vijverberg & Van den Bercken, 1990; Bogh et al., 1998), and c) causing mosquitoes to
leave treated nets or avoid approaching treated nets (Lindsay et al., 1991; Miller et al.,
1991; Rutledge et al., 1999), or even avoiding entering rooms or houses containing ITNs
(Gimnig et al., 2003b; Mathenge et al., 2001). Evidence is mounting that, under dense
deployment, ITNs can offer some protection to nearby persons not actually under a
bednet (Gimnig et al., 2003a & b). However, the degree to which the various alternative
explanations above are contributing to such positive end results has been unclear and the
subject of some debate (Habluetzel et a1, 1997; Roberts et al., 2002). Definitive answers
to this question could inform managers of vector-control programs about how the ITN
tactic might be optimized as well as sustained by avoidance of resistance, be it metabolic

or behavioral.

24

The current study, employing modern techniques for quantifying animal behavior
in high detail, was part of a research program whose objective was to quantify the relative
contributions of the above possible mechanisms to permethrin- and deltamethrin-
impregnated ITN mode-of—action against Anopheles gambiae Giles, the major vector of
malaria in Sub-Saharan Africa. Here we report results of laboratory studies conducted at
a small spatial scale, in: a) cages fashioned from netting and experiencing little air
movement (static), and b) a small wind-tunnel where air-flow was controlled through

treated and un-treated netting. Results at the house scale will follow in Chapter 3.

MATERIALS AND METHODS

Mosquitoes

The Kisumu laboratory strain of An. gambiae 5.5. was used throughout. Mated-
female stock for this culture was house-collected between 1990 and 1992 from multiple
sites near the Kenya Medical Research Institute (KEMRI) at Kisian (V ulule et a1, 1994).
Eggs were deposited on moist paper at 28 1 10C under a photoperiod LD12112 h. After 1
day of incubation, they were transferred to plastic containers measuring 27 x 19.4 x 9.5
cm and holding 700 - 800 ml of tap water for larval rearing. Most of the lid area of the
rearing containers was covered with 0.25 mm polyester nylon mesh (American Home and
Habitat, Squires, MO, U.S.A.). Light at approximately 700 lux was provided at the water
surface by fluorescent bulbs (Lumichrome 40 W, 122 cm T12, 5000K, CRI96, M & M
Lighting Co., Ronan, MT, U.S.A.) mounted directly over the rearing containers. Larvae
were reared on ground and sieved (100 mesh) food mixture (2:1) fish food (TetraMin(R),
Blacksburg, VA, USA): brewers' yeast (Bio-Serve). Amounts served were: 5 mg on day

1, 28 mg on day 2, 36 mg on day 3, 120 mg on day 4, 200 mg on day 5 and subsequent

25

days before pupation. Pupae were collected from the containers each day and transferred
into mosquito pupal breeders (21 cm height 12 cm diameter) consisting of 2-L clear
styrene containers (BioQuip Products, Gardena, CA, U.S.A.). The bottom portion of each
pupal breeder held 250 ml water with several hundred pupae. A plastic lid between the
two portions contained a clear, vinyl funnel through which emerging adults flew into the
upper portion ventilated at the top with aluminium screening (5 cm diameter). Eclosed
adults were transferred into rearing and supply cages.

Adult female mosquitoes were fed via an artificial blood-feeder (Huang et al.,
2005) consisting of a water bath, fountain pump, Tygon(R) tubing (Saint-Gobain
Performance Plastics, Beaverton, MI, USA), and blood reservoir was used. The
reservoir was a metal jar lid (6 cm diameter x cm height) hot—glued to a flattened copper
tube connected to Tygon tubing. Warm water (44°C) was pumped from the water bath
through the tubing (1.8 cm diameter) and through the copper tube, to heat the blood in the
reservoir to 36°C. At each feeding, 15 m1 of defibrinated horse blood (Lampire
Biological Laboratories, Pipersville, PA, USA.) was poured into the feeder reservoir and
sealed with two layers of very thinly stretched ParafilmTM. The feeder was placed
membrane-down on top of the rearing cage, arranged so its top was flat. A small piece of
dry ice and four paper towel strips (2 x 25 cm) soaked in Lirnburger cheese solution (23%
w/w, stored overnight at 4°C) were placed near the blood reservoir as attractants and
stimulants (Knols et al., 1997). Mosquitoes were allowed to engorge for 1-2 h every other
day. Sugar feeders, comprised of a cotton wick connected to a 20 ml reservoir of 10%

honey solution, were continuously present on the cage floor (Huang et al., 2005).

26

Experimental Series I — Tests of Insecticide- Treated and Untreated Netting in Small
Static Cages

These tests were conducted in 20 cm diam. x 20 cm tall cylindrical cages (Fig 2.1).
The cage top and sides were constructed of standard polyester lOO-denier 156-mesh bed
netting material (Siam Dutch Ltd, Thailand) sewn together and attached with VelcroTM to
external wire framing for support. The floor was clear plastic, a substrate upon which An.
gambiae do not normally prefer to sit. Mosquitoes were inserted or withdrawn from cages
via an aspirator inserted through a small slit in a fabric seam, sealed with Velcro”. Three
types of static cages were tested in a walk-in environmental chamber (12L: 12D, 28°C,
RH 75-85%): 1) all untreated fabric, 2) all ITN fabric, and 3) half-untreated and half-
ITN. An individual female mosquito was introduced into an arena during the last 2 h of
photophase and observed continuously for 30 min, a period sufficient for evaluation of
ITN efficacy (Hougard et al., 2002). Using a lap-top computer operating Version 5.0 of
the Observer (Noldus, http://www.noldus.com) behavioral quantification software, the
following data were recorded: frequency and duration of flight; number of contacts with
cage walls; time and location of landings; and elapsed time sitting and walking on cage
surfaces. Also recorded were timing and duration of behaviors associated with toxicosis:
postural changes; tremors; leg autotomy (Moore and Tabashnik 1989a, b); knockdown;
and loss of responsiveness to tactile stimulation. Mosquitoes in advanced toxicosis were
placed in 10 ml glass vials and monitored for time of death or recovery. Test A of this
series compared an untreated cage to fully-treated PermanetTM (Vestegaard Frandsen;
55mg wash-proof deltamethrin/m2 impregnated into polyester) (WHO, 2000), Olyset

(Sumitomo; 1 g Permethrin/m2 of polyethylene fiber (N ’Guessan et al., 2001), and

27

Netprotect (Intelligent Insect Control) cages (10 replicates of each); Test B compared an
untreated cage with half-treated Permanet, Olyset, and Netprotect cages (10 replicates of
each). The experimental design for treatments within a test was randomized complete

block; replicate blocks were accumulated over time.

 

 

 

 

 

 

 

 

Velcro—v! 7 ll
Wire -—>~
Framing S 20 cm
eam —>
Polyester
Netting
C
age 20 cm II
Untreated
Plastic Floor

Figure 2.1. Representation of a 20 cm diam. x 20 cm tall static cage used for
observations of Anopheles. gambiae females responding to lTNs. Cages were
either constructed: 1) entirely of untreated polyester netting, 2) of half-
untreated netting and the other half of Permanet, or 3) half-untreated netting

and the other half Olyset. Halved cages were split vertically.

28

Experimental Series 2 — Tests of Untreated and Insecticide- Treated Netting in a Wind
Tunnel with Host Cues

A wind-tunnel (Figure 2.2) suitable for quantifying the behaviors of flying
mosquitoes in the vicinity of insecticide-treated vs. control screens was constructed after
the designs of Miller and Roelofs (1978), Gillies and Wilkes (1981), and Miller and
Gibson (1994). A variable-speed electric blower was joined by flexible plastic sheeting to
a wind-laminator, consisting of a 60 cm high x 70 cm wide x 15 cm deep wooden frame
having 3 layers of stretched sheer fabric (ca. 10 mesh/cm) stapled to each face. Attached

in series to the laminator were two 90 cm long x 70 cm wide x 60 cm high chambers of

   
   

Netting Insert

 

Exhaust Fan

Supply Fan

Figure 2.2. Wind tunnel for testing responses of Anopheles gambiae
females to insecticide-treated bednet materials placed downwind of host

CUES.

29

clear PlexiglasTM (see Figure 2.2), joined by a wooden frame slotted so as to permit
insertion of exchangeable panels of bet-netting, stretched and supported by independent
frames. Types of netting panels used were: untreated netting, all Permanet'm, all
Olyset'm, or half-untreated joined to half-treated deltamethrin (25 mg/m2) or permethrin
(500mg/m2) panels with a vertical seam at their midpoint. Chamber 1 housed sources of
host cues released as one centered set when a panel of netting was homogeneous, or as
two sets centered on each half-panel when two types of netting were tested
simultaneously. Heat approximating that from human skin was provided by mounting a 4
cm diam hexagonal head of a 7 cm long x 2.5 cm diam steel bolt against the upwind side
of the fabric panel. A resistive-wire wrapping controlled by a variable transformer heated
the bolt to 31°C. A stream of air flowing over an aqueous slurry of limburger cheese
(Knols et al., 1997) and infused with C02 released from a pressurized tank fitted with a 2-
stage regulator, was directed through Tygon tubing, 3.2 mm (i.d.) so it flowed over the
warm bolt. The C02 release rate at ca. 250 ml/min, which approximates that produced by
a typical human subject, was controlled by using Linde precision control valves and
measured using calibrated Linde rotameter-type flowmeters.

Chamber 2 was the working section of this wind-tunnel. The roof and distal end of
this section were fitted with untreated polyester netting to provide suitable perching sites
for An. gambiae females under test. Its downwind end was closed with metal window
screening, in which openings were made to receive 11 cm long x 5 cm diam plastic cages
with removable screened ends, for releasing individual female mosquitoes (release cages)
within a stimulus plume. An air-scoop connected to a second blower fan exhausted all

stimulus fumes from the laboratory. Beneath the floor of Chamber 2 was white paper on

30

which fifty 4-cm diam black circles were randomly positioned to provide a strong
stimulus for optimotor responses (Knols et al., 1994). Wind flowed laminarly through the
working section at 0.2 m/s when all hinged doors permitting access to the chambers were
closed. The flight chamber was illuminated by a 40W fluorescent light bulb suspended
above the tunnel. Light was diffused through layer of 3 mm opaque, white acrylate and
another of white filter paper reduced light intensity to 6.5 i 1.5 lux in the chamber. No
other illumination was present in the room.
Wind tunnel procedure

Individual 2-3 day-old female An. gambiae mosquitoes reared as above and.
having previous access to sugar solution but not blood were gently transferred by
aspirator into a release cage. Cages were positioned in clean air on the floor of the wind-
tunnel working-section for 5 - 10 min to allow for acclimation to the flowing wind. To
initiate a run, a release cage was inserted through the end-screen and into the plume of
cues; the upwind end of the release cage was then gently removed. Continuously for the
next 30 min, a researcher sitting beside the working-section of the wind-tunnel keyed the
following behaviors into a laptop computer running ObserverTM software: sitting
(stationary state with at least two pairs of legs in contact with surface); walking (forward
or sideway alternate movements of legs on surface); flying (free airborne movements on
the wing); net-flying (flying while mouthparts and front legs kept in near constant contact
with the net); hopping (short leaps of flight interspersed with short periods of sitting);
labella tapping (occasional pressing of mouthparts on net surface and through openings in
the net ); leg lifting (alternating rapid movements on and off the net or other surface

while in a sitting position); grooming (rubbing/cleaning legs, wings and mouthparts using

31

a pair of legs); antennae raising (elevating antennae beyond normal position); jerking
(twitching motions of whole body or parts of it), leg dropping ( legs detaching from main
body), erratic movements (staggering rapid motions of flying, walking and sitting),
knockdown (loss of normal posture and locomotion) and wing spread (wings remain
outstretched while sitting or in knockdown state). The position where each behavior
occurred in the arena was noted and recorded as: Upwind treated or untreated (section of
a treated or untreated screen in the netting insert); Downwind mesh (metal window
screening at downwind end); Netted side (vertical wall side of Chamber 2 opposite an
observer fitted with untreated netting); Netted roof (ceiling side of chamber 2 fitted with
untreated netting); Floor (bottom area of Chamber 2); Front plexiglas (immediate
transparent vertical side facing observer); Treated/Untreated heat (nut heads placed
behind treated or untreated screen, respectively); Intersection (exposed frame positions
and other locations not defined above). After the 30 min, the specimen was placed into a
holding cup for 24 h with access to 10% sugar solution and water, for assessment of leg
autotomy and mortality. A total of 25 replicates was accumulated for each type of net-
treatrnent to be reported. Tests were blocked so that all ITN types were successively
tested in random order at one setting.
Data Analysis

Before analysis, sets of data were examined under the Descriptive Statistics/Basic
Statistics feature of Systat for levels of kurtosis and skewness. Data were transformed as

1/2

necessary, usually to (x + 0.5) , and occasionally to logm(x + 0.5). ANOVA was

performed using SAS version 9.1 (2000) or SYSTAT Version 10 and means were

separated using Fisher’s LSD test. Time-event tables were generated within Observer and

32

various plots of the data were examined to identify outcomes justifying statistical
analyses. Lag sequential analysis was carried out to visualize the temporal structure of
sequence of events. It allowed for calculation of frequencies of transitions between pairs
of events within a certain lag in a time series. The first event of each pair was designated
the criterion event and the second as the target event. The lag was therefore the time
separating the criterion and target events.

Transition matrices of pro-knockdown and post—knockdown behaviors, with the
diagonals treated as logical zeroes (no behavior could follow itself), were constructed as
per Walker and Archer (1988). Only analyzable first order transitions were considered. A
2 x 2 contingency table analysis was conducted on each cell to test that the before/after
transition frequency differed from chance. The main point of interest was whether the
transition was higher than expected. Kinematic graphs were helpfirl in visualizing
behavioral transitions and frequencies in untreated and treated-net scenarios and in pre-
knockdown and post-knockdown periods.

RESULTS
Experimental Series 1 -— Tests of Insecticide- Treated and Untreated Netting in Small
Static Cages

Test A — Untreated cage vs. [ulna-treated Permanet, Olyset, and Negemtect cages

The average times An. gambiae allocated to various behaviors during the

 

combined 30-min observations while interacting with untreated and fully-treated
Permanet, Olyset, and Netprotect ITN static cages are shown in Figure 2.3. On average,
the duration was 12% for flying, 82 % for sitting, and 6% for walking on the wall, roof,

and floor of an untreated control (Figure 2.3). Flying within fully-treated Permanet,

33

 

 

 

 

 

 

100%
‘E
o
'g 80%
Z I leg dropping
g 60% I Knockdown
'6 I walking
E, 40% I sitting
"E a flying
3 20%
1:
g 0%

control Netprotect Olyset Permanet
Static cage

Figure 2.3. Average percent time Anopheles gambiae allocated to various behaviors
during the 30-min observations of single females in untreated static cages or cages

fully-treated with deltamethrin (Netprotect and Permanet) and permethrin (Olyset).

Olyset, and Netprotect ITN cages constituted 11, 12, and 10% of observation time,
respectively. A mean of 80% of the time was spent on the walls of the untreated cage
(Figure 2.3). However, sitting duration was 37, 28 and 18% of the time for the fully-
treated Permanet, Olyset, and Netprotect ITN cages, respectively. Time spent on the floor
for An. gambiae females was 63, 62 and 60% for fully-treated Permanet, Olyset, and
Netprotect ITN cages, respectively (Figure 2.4). However, average time allocations are
of less interest behaviorally than shifts in behaviors over time.

Mosquitoes in control cages were consistent in their behaviors through time. No
statistical differences across time were found for: types of behavior (Figure 2.5 A, C, E),
durations of behavior (Figure 2.5 B, D, F), positions of behaviors (Figure 2.6 A, C, E),

and durations of behaviors at specified positions (Figure 2.6, B, D, F). However, presence

34

 

 

 

 

 

 

 

 

 

 

.100%
S
E 80%1
'8 60%— Dfloor
\° amof
2’ 40%J IVldI
.8
S 20%l
3: 0%-

oortrol hbtpnotect Oyset Pemrenet

Lowtionincage

Figure 2.4. Total time spent in given locations in cage during 30-min
observations of Anopheles gambiae females in static cages treated with three

types of ITN fabric

of insecticides caused major shifts in behaviors through time. Initially, all ITNs
significantly elevated frequency (Figure 2.5 A) and duration (Figure 2.5 B) of flying.
Flight durations under Permanet exposure were not as high as those for Netprotect and
Olyset. Also elevated initially was frequency (Figure 2.5 C) but not duration (Figure 2.5
D) of sitting events. Since every flight ends with a new sitting event, it is likely that
sitting was elevated only because it must follow flight. This is why Figure 2.5 A is
virtually identical with Figure 2.5 C. The mean frequency (Figure 2.5 E) and mean

duration (Figure 2.5 F) of walking were never elevated by insecticide exposure.

35

Mmddmm 0 Mean Frequency of Flying >

MFmlealldng m

 

 

..5
0|
0

-o— Netprotect‘
+ Olyset
f": Bimflli

.s
N
o

   
 

(a) O) CD
0 O O
O’
1

 

 

 

 

 

 

 

 

420

 

 

300 ~
240 —
180 -
120 ~ DB

§
91;]

O
l
1 or
a:
Mean Duration of sitting 0 Mean Duration of Flying (D
g o
g y
0'

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Quarter Quarter

Figure 2.5. Mean total frequencies and mean total durations for behavioral
responses of Anopheles. gambiae females broken out per quarter of the 30
min observation period. This test was conducted using either fully-untreated
cage walls and roof or fully-treated cage walls and roof. Data within a column
sharing the same lower-case letter are not statistically significant; data

across quarters sharing the same capital letter are not statistically different.

36

 

5 _’ 'A_ * — ‘- '_ “_:_:—7..:_;T__
I a I w—Control

I «gr-NetProtect
--Olyset
sis-Permanet

a a A l

g b a l

-—l,

0 .
C 1 2 3 4
1

   

 

on the Wall

 

 

 

 

 

v

8

x NS
= d

O

 

 

 

Mean Frequency of Arrival Mean Frequency of Arrival

 

 

Mean Frequency of Arrival
on Floor

 

Quarter

 

Vlalt on Wall

Mean Duration on Roof Mean Duration of Stay per

Mean Duration on Floor

 

480
400-
320~
240‘
160i

80 ~

 

 

 

 

Oo

 

50‘
40‘
30“
201
10*

 

480

 

 

400-
320-
240-
160‘

80 -

 

a8
a8

 

 

Quarter

Figure 2.6. Mean total frequency of arrival at particular positions and mean

total durations of stay at those locations for female Anopheles gambiae in

static cages, as broken out by quarter of the 30 min observation period. This

test was conducted using either fully-untreated cage walls and roofor fully-

treated cage walls and roof. Data within a column sharing the same lower-

case letter are not statistically significant; data across quarters sharing the

same capital letter are not statistically different.

37

The mean elapsed time for first knockdown for Netprotect, Olyset, and Permanet
are given in Table 2.1. Elapsed time for first knockdown by Netprotect was significantly
shorter than that for Permanet. Olyset fell between and was not significantly different
from Permanet and Netprotect. At about the time of first knockdown, the stimulatory
effects of the lTNs begin to diminish as frequency of erratic movements increased
(Figure 2.7 A). Also at this time labella tapping and grooming frequencies diminished.
Initially labella tapping had been elevated for Olyset and grooming had been depressed
by Permanet and Olyset. By quarter 3, flying, sitting, and walking were all significantly
reduced below levels for mosquitoes in control cages (Figure 2.5). Knockdown (Figure
2.8) shifted mosquito presence from the walls to the floor (Figure 2.6 B vs. F). By the
final quarter of the test, virtually all mosquitoes were lying on the floor, incapable of
flying, walking or sitting (Figure 2.5 A, E, C).

Test L— Untregted ca e vs. hal -treated Permanet 0l set and Net rotect ca es

Figure 2.9 shows the average times during 30-min observations that An. gambiae
allocated to various behaviors while interacting with untreated small static cages and
half-treated Permanet, Olyset and Netprotect ITN cages. Flying within half-treated
Permanet, Olyset, and Netprotect ITN cages constituted 15, 19 and 10% of observation
time, respectively. Sitting occupied 48, 46, and 45% of the time for the half-treated
Permanet, Olyset, and Netprotect ITN cages, respectively. The percentage of time An.
gambiae sat on the floor of half-treated Permanet, Olyset, and Netprotect cages vs. the
control cage was 18, 20, and 19 vs. 46%, respectively, (Figure 2.10). Consistencies and

changes in behaviors in these cages during 30-min observations follow.

38

Table 2.1. Comparison of elapsed time for first knockdown event for Anopheles

gambiae females in small cages fully— or half-treated with ITN fabrics. Means

followed by a common letter at any place in the table are not significantly different

at p = 0.05.

 

Cage Treatment

Elapsed Time (Min :1: SEM) for First Knockdown

 

Netprotect Olyset Permanet

 

Untreated Netting
Fully-treated Cage

Half—treated Cage

No knockdown No knockdown No knockdown
8.1:t1.0c 11.711.3bc 14.811.5ab

14.7:l:1.2 ab 14.1 :l:1.2 ab 17.8:133

 

39

Frequency of Labella

Frequency of grooming

 

 

 

 

 

 

 

 

 

 

 

A ':EothroT—I
1O - aA ' :5 3A —o—Netprotectl
8 ‘ a x l—a—Olyset ‘
6 - I:><:Pertn_a_n§t_l
4 ‘ E
. a
2 b Bl p p a B
O c v v -
1 2 3 4
A3
9—
4

 

 

 

 

 

 

 

Quarter

Figure 2.7. Mean total frequencies for behavioral responses of Anopheles

gambiae females broken out per quarter of the 30 min observation period.

This test was conducted using either fully-untreated cage walls and roof or

fully-treated cage walls and roof. Data within a column sharing the same

lower-case letter are not statistically significant; data across quarters sharing

the same capital letter are not statistically different.

40

 

2.0 p
|+Control (1

     
  

 

 

 

 

:
g a B l—r3—Netprotect
1: 1.5 1 X l
f, a B +Oly5et i
5 —x— Permanet |
“a 1.0 l \
>. a " a B
0
5
g 0.5 ,
2
“ b A b b b

0.0 ' 3 f c v 3

1 2 3 4
Quarter

Figure 2.8. Mean total frequencies for knockdown of Anopheles gambiae
females broken out per quarter of the 30 min observation period. This test was
conducted using either fully-untreated cage walls and roof or fully-treated
cage walls and roof. Data within a column sharing the same lower-case letter
are not statistically significant; data across quarters sharing the same capital

letter are not statistically different.

41

100%

 

     

 

 

 

E
.9
‘5
E &1%
§ Ilegdomirn
"6 60% nKnodrdom
2‘: avdking
5 40% Isittirg
.3 .
g UIIyII'Q
3 20%
E
O
'- 0%

oontrol taprotect Oyset Pemmet

Staticcage

Figure 2.9. Average time Anopheles gambiae allocated to various behaviors during
the 30-min observations of single females in untreated static cages or cages half-

treated with deltamethrin (Netprotect and Permanet) and permethrin (Olyset).

The same consistencies reported above for mosquitoes in control cages were
again evident (Figure 2.1 1). Moreover, the repeatability both for controls and ITN
treatments between these full-treated and half-treated cage experiments was striking.
Figures 2.11 and 2.5 are a near match when overlaid. Because cages in the latter
experiment were only half-treated with insecticide-treated netting, there was less
consistency in the timing in the first insecticide exposure. For example, a female landing
and sitting on untreated netting would absorb no insecticide until leaving and alighting on
treated netting. Thus, it is not surprising that elapsed times for first knockdown were
delayed for half-treated cages relative to fully-treated cages (Table 2.1).

Although significant differences were not found between half and fully—treated

Permanet and Olyset cages when each was analyzed separately, statistical significance

42

1GJ% -

 

 

 

 

 

8

E EP/r

,g 60%- aneuralloor
é Dutreaalsicb
g 40% ltrededside
1:3 20%1

i on..-

oortrol handed Oysa Pem'ma
Staticmge

Figure 2.10. Average times Anopheles gambiae females spent in particular
locations in untreated vs. half-treated static cages as averaged over the whole

30 min observation period.

was obtained at P = 0.03 when paired data for all three net types were combined. It
follows that the depressions in frequencies and durations of behaviors seen in Figure 2.11
slightly lag those in Figure 2.5. Overall, the main effect of the insecticides was to
stimulate flying until knockdown commenced. The behavioral repertoire became
progressively dominated by erratic movements (Figure 2.14) and knockdown events
(Figure 2.15), until the females accumulated on the cage floor as knockdown became
permanent.

No evidence was found that An. gambiae females detected and avoided
insecticide-treated sections of half-treated cages. This is shown in Figure 2.12, where the

profiles for treated and untreated cage halves were statistically identical for all cages

43

Frequency of Fly

F reqency of Sitting

Frequency of Walking

 

 

 
 
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

50 _ __ -- 150
Ia A A +Control—1 A
40 - -o— Netprotect 3 120 ‘
I+Olyset g l
30 ’ 1:“:PfflLmel I}; 90
20 y 8 60i
b a E
10 - 5 3° ‘
b B
o T . — 0
1 2 3 4
50 IaAIR C A :23 .
4o 2 3”
"22’ 300 4
3° “ '5 240 l
a ‘0-
20 b a B a g 180 -
w « E ‘23 i
b b a
0 r f r j 0
1 2 3 4 1 2 3 4
10 60
8’
'2 4O 4
g a
.._ 30 ‘ ,/’-/\
o a
.5 20 ~ air a
g 10 — b B
b B
0 fl 2 - _
1 2 3 4 1 2 3 4
Quarter Quarter

Figure 2.11. Mean total frequencies and mean total durations for behavioral
responses of Anopheles gambiae females broken out per quarter of the 30
min observation period. This test was conducted using either fully-untreated
cage walls and roof or half-treated cage walls and roof. Data within a column
sharing the same lower-case letter are not statistically significant; data

across quarters sharing the same capital letter are not statistically different.

44

 

 

 

 

 

 

 

 

1O ..flwuw‘w
Quarter1 f o Treated ]
8 a ;_0 __U_ntLe_=1te9.
6
0'3
0 4 -‘ °l b C,ab
E 2 * b 0 D a
a)
2:. 0 -
Ti 6* e‘} a}
< 90 Q0
“5
5‘ 10
g Quarter3
:3 8 ~
0'
a:
E 6 -
g 4 _ NS
0) O
2 a
2 ‘ a
D 9
O I
6* 6‘
’6 x, 9 0
000 do 0% 0&0
e° ‘2

Half-treated Cage

 

 

 

 

 

 

 

 

10
Quarter
8 ..
6 ‘ 9 a
4 u
o ab 9 ab
2 4 b o
0 r
e Qt} 6‘ e“
<15 e s" 0“
0° « «06°
10
Quarter4
8 _
6
4 _
2 ~ NS
0 D
o : . 0 fl 9 . °
e a ‘9‘ is
06“ Q<°‘ o\* 66”
#5 Q°
Half-treated Cage

Figure 2.12. Mean total frequencies of arrival per side by Anopheles gambiae

females broken out per quarter of the 30 min observation period. This test was

conducted using either fully-untreated cage walls and roof or half-treated cage

walls and roof. Data within a column sharing the same lower-case letter are

not statistically significant.

45

containing insecticide. This was true throughout all four quarters of the 30-min tests. In
quarters 1 and 2, Olyset elevated frequency of arrival on both untreated and treated cage
walls over the values obtained for fully-untreated cages (Figure 2.12 A and B). This
effect disappeared in quarters 3 and 4 (Figure 2.12 C and D). Durations of stay on
untreated vs. treated halves of half-treated cages were identical across all 4 quarters and

for all three ITN types (Figure 2.13).

46

Mean duration per stay

Mean duration per stay

 

 

 

 

 

 

 

420 rm“
350 - Quarter1 lo Treated 1.
FLUDL'EatEQi
280 — D
210 ‘ I; a .
140 -
7o — NS
0
\ 6. x 3;
6° 0 a? e
<‘ 0‘ ~\
0° eggs (2669
420
360 . a Quarter 3
300 4
240 -
180 -
120 r b D b
60 4 C] b D
0 Ab ° b . b
\ x. \
066° ‘oxelé' dtx 600(3’
0 eéa Q0
Half-treated Cage

 

Mean duration per stay

Mean during per stay

 

420
350 *
280 ‘
210 e
140 ‘

 

Quarter 2

 

420

 

350 ~
280 ~
210 ~
140 e
70 e

 

a Quarter 4

 

Qe
Half-treated Cage

Figure 2.13. Mean total duration of stay by Anopheles gambiae females on

treated vs. untreated sides of half-treated cages broken out per quarter of the 30

min observation period. Data within a column sharing the same lower-case letter

are not statistically significant.

47

 

 

 

  
 
  

 

 

 

 

 

 

 

° :
>0 +Nelprolect
2 E 10 . +0lysel
g > l-x-Permanet
o — — —— 1
32 x
02.2 5 ‘ a
c“
a 2
2" = a b b
0 4 e e t
1 2 3 4
Quarter
.6 1.0
U
3' s
C Q
0 Q
3 N
on“
2 2
h —
c 3 NS
2 2
5 0,0 g < : \, x/:
1 2 3 4
Quarter

Mean duration oferratic

Mean duration of labella

200

.3
0'1
0

movement

1.5i

tapping

0.5 ~

0
O

01
O

 

 

O

 

 

 

 

 

 

a b
:: T 3 fl—'
1 2 3 4
Quarter
NS
1 2 3 4
Quarter

Figure 2.14. Mean total frequency and duration of behaviors by Anopheles

gambiae females on treated vs. untreated sides of half-treated cages broken out per

quarter of the 30 min observation period. Data within a column sharing the same

lower-case letter are not statistically significant.

48

 

—L

 

 

 

 

 

g x
3 0.8
:8: a8 r—fi* __ 1
a: 0 6 l+Contol 3
E ' i —o—Netprotecti
3 I—.-—Ol et ‘
g 0.4 . I ys }
:- L:x- Permanetl
r:
c 0.2
8
a 0 f

4

 

Quarter

Figure 2.15. Mean total frequency of knockdowns by Anopheles gambiae females
on treated vs. untreated sides of half-treated cages broken out per quarter of the 30
min observation period. Data within a column sharing the same lower-case letter
are not statistically significant; data across quarters sharing the same capital letter

are not significantly different.

Experimental Series 2 -— Tests of Untreated and Insecticide- Treated Netting in (1 Wind

Tunnel with Host Cues

Test A — Untreated cage vs. [ullz-treated Permanet, and Olyset screens

The average times An. gambiae allocated to various behaviors during the 30-min
observations when exposed to untreated and fully-treated Permanet, and Olyset ITN
screens at the upwind end of the wind tunnel are shown in Figure 2.16. When no
insecticide was present, 74% of the time was spent sitting (Figure 2.17), of which 50%
occurred on the upwind screen (Figure 2.17). When the upwind screen contained
Permanet, females on average spent 48% of the total time sitting, of which 26% occurred

on the upwind screen. When Olyset was present, females spent 42% of the total time

49

 

   

 

’E 100%

o

'5 I rbppirs

S 80% El Enatic movt.
0 60°/ I Lmella taaping
‘6 ° a net flying

.\°

= 40% ""'°°'_ “1‘3“"
'3 vralkrng

g . .

'3 20%, I Slt‘tlng

E I] Flying

19 0%

 

 

 

Oyset

Wind Tunnel
Figure 2.16. Average time Anopheles gambiae allocated to various behaviors
during the 30-min observations of single females in a wind tunnel with untreated
and fully ITN-treated upwind screens, as averaged over the full 30-min

observational period.

sitting, of which 28% was spent on the upwind screen. The overall pattern in time budget
within the wind tunnel was more similar to that within the half-treated static cage (Figure
2.9). Duration of stay on the floor for Permanet and Olyset was double that of stay on the
floor for untreated screen (Figure 2.17). The behaviors expressed by An. gambiae females
in the wind tunnel under no exposure to insecticide-treated netting were less consistent

through time than were behaviors in static cages.

50

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

,8

81°10 alrtersections

IFra‘tfledga
.5 60%4 [Floor
32 oNetredmor
‘8' 40%4 nNertedsrde
E IWrescreen
'6 20%” chreetedsoreen
3 0% i . e
Cortrol Perrra'ra Oyset
Wind Tunnel

Figure 2.17. Average times Anopheles gambiae spent in particular locations
during the 30-min observations of single females in a wind tunnel with
untreated and fully ITN-treated upwind screens, as averaged over the full 30-

min observational period.

Frequency and duration of flying by control mosquitoes in the wind tunnel tended
to rise over time (Figure 2.18 A, B). The rise through time was significant for mean
frequency of flying (Figure 2.18 A), and close to significant (P = 0.057) for mean
frequency of walking (Figure 2.18 E). An upward tendency through time was observed

for various other behaviors by control insects (Figure 2.19; Figure 2.20).

As in the static cages, insecticide-treated fabrics at the upwind end of the wind
tunnel elevated some behaviors of exposed An. gambiae. Both Olyset and Permanet
significantly increased frequency (Figure 2.18 A) and duration (Figure 2.18 B) of flying;

and, Olyset was significantly more stimulating than was Permanet. Thus, both Olyset and

51

Permanet functioned as locomotor stimulants (Figure 1.6). However, these insecticides

did not elevate walking (Figure 2.18 E) or labella tapping (Figure 2.19).

Mean elapsed time until first arrival on the upwind screen of the wind tunnel for
control, Permanet, and Olyset was 4.2 i 1.4, 3.4 :l: 1.1, and 3.2 i 0.8 min, respectively;
differences not significant by AN OVA on data transformed to (x + 0.5)'/' (P = 0.98). This
suggests there was no repellent effect of these insecticides, i.e. no tendency to fly
downwind when exposed to air flowing over insecticide. Mean elapsed times for the first
knockdown event by Olyset and Permanet in the wind tunnel for the experiment were
18.5 d: 1.2 and 17.3 d: 1.4 min, respectively, differences not significant. These values
were only slightly greater than for elapsed time until the first knockdown in half-treated
static cages (Table 2.1). Perhaps the presence of moving air and host cues in the wind
tunnel led to speedy delivery of a lethal dose of insecticide in the wind tunnel, even

though it was much larger than a static cage.

Beginning in the second quarter and until the end of the tests, the frequency and
duration of erratic movements increased dramatically (Figure 2.19 E, F), and first
knockdown soon followed early in the third quarter. Thereafter, behaviors such as flying
(Figure 2.18 A,B), sitting (Figure 2.18 CD), and walking (Figure 2.18 E,F), labella
tapping (Figure 2.19 A, B), and net flying (Figure 2.19 C, D) all diminished progressively
until 90 and 80% of females exposed to Olyset and Permanet, respectively, were
continuously confined to the floor, beginning at 9.7 i 0.9 vs. 3.4 :l: 1.4 min, respectively

(difference significant at P = 0.035 by ANOVA on untransformed data).

52

 

 

150

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

40 + °"'° c 120 ~
"5 —°—Olyset "6
g 30 “ Left—”Permanet) 5 90 '
2° “ g 60 -
E E, U
10 - a "S 30 —
g b ng g
0 l r r 0
1 2 3 4
E 50 g, 420
.32 c ._
"’ 4o 4 a 360 l
“5 g 3004
g 30 ” l a a 5 2404
20 — a a E 180*
8 ab 3
._ 10 - '0 120 a
r: b g -
8 b b 60
2 0 ‘ 5 0 e
1 2 3 4 1 2 3 4
.8 1° .8’ 60
E 8 4 E i 50 ‘ F
"5 '5 40 .
6 -r
4 “ 5 20
3 8
4: 2 - c 10 . a a b
g 0 g 0 _b8
1 2 3 4 1 2 3 4
Quarter Quarter

Figure 2.18. Mean total frequencies and mean total durations for behavioral
responses of Anopheles gambiae females broken out per quarter of the 30
min observation period. This test was conducted in a wind tunnel using either
fully untreated or fully ITN-treated upwind screens. Data within a column
sharing the same lower-case letter are not statistically significant; data across

quarters sharing the same capital letter are not statistically different.

53

Mean frequency of
labella tapping

Mean frequency of
erratic movement

Figure 2.19. Mean total frequencies and mean total durations for behavioral
responses of Anopheles gambiae females broken out per quarter of the 30
min observation period. This test was conducted in a wind tunnel using either
fully untreated or fully ITN-treated upwind screens. Data within a column
sharing the same lower-case letter are not statistically significant; data across

quarters sharing the same capital letter are not statistically different.

Mean frequency of net
flying

—I
O

..s
O

 

 

ONhQG

can;
1 4» Olyset

$999139

 

 

 

Mean duration of
labella tapping (s)

‘8)“ng

O

 

 

 

 

 

 

Mean duration of
erratic movement (3)

 

 

 

54

Mean duration of net
flying is)
8

O

 

O

CO

 

O

 

 

150

120 ~

CD
0

8

 

 

NS

 

 

_a_a
N
o

3333

 

 

 

Quarter

 

 

 

Working in a wind tunnel where air flow was controlled, enhanced ability to
characterize the behavioral mechanisms mediating ITN effects in the terminology of
Figure 1.6. It has already been noted that neither Permanet nor Olyset diminished the
frequency of flying upwind in air flowing over these insecticides. Thus the Perrnethrin
and Deltamethrin dosages used here were not repellents i.e., repellents cause responders
to move downwind or out of stimulus plumes by maneuvers always directed away from

the source of stimulation.

Frequency of arrivals on the upwind screen of the wind tunnel was significantly
elevated for Olyset over the control (Figure 2.20), as would be expected for a Locomotor
Stimulant (Figure 1.6). But, the rate of stay on the treated screen was not reduced relative
to that for the untreated screen (Figure 2.20). Rather, mosquitoes stayed somewhat longer
(although not significantly so (Figure 2.21)) than their counterparts on an untreated
screen. Nor can Olyset be considered a Deterrent (Figure 1.6) in the present context, as
the frequency of visitation to the Olyset source increased slightly from quarter 1 to
quarter 2, rather than decreased relative to the visitation rate to the control (Figure 2.20).
Thus by the terminology of Figure 1.6, Olyset can be termed locomotor stimulant.
Permanet was likewise documented to be a locomotor stimulant because it increased
flying frequency in the wind tunnel (Figure 2.18 A). However, this effect by Permanet
was not sufficiently strong so as to significantly increase visitation rate to its source on
the upward screen (Figure 2.20). Like Olyset, Permanet proved to be only a locomotor
stimulant, albeit a weaker one than Olyset. In quarters 3 and 4, visitation rates to the
upwind screens and duration of stay (Figure 2.21) diminished as mosquitoes became

increasingly poisoned. Frequencies of visitation and duration of stay on the downwind

55

 

 

 

 

 

 

E 10
o .4— _.______.
g {:0— Control _}
§ 8 1 1+ Olyset
3' itffiflléflefl
S 6 l
1’
é
a 4 ~ 8
I-
: ab bed
3 2 a bed Ar,» be A de
3 Cd bc fi
a- cd
IE ode e
0 . . .
1 2 3 4
Quarter

Figure 2.20. Mean total frequencies of arrival by Anopheles gambiae females
on the upwind screen of a wind tunnel, broken out per quarter of the 30
min observation period. This test was conducted in a wind tunnel using
either fully untreated or fully ITN-treated upwind screens. Data within a

column sharing the same lower-case letter are not statistically significant.

 

 

 

 

 

 

 

E
g 420
«3 350. +Control '
E —o—Olyset
g, 300 4 a +Permanetl
3 240 ab
g ab
5 180- abc
“a bcdl ”“1
g 1201 d de
g C de
a 50 e
5
g 0 .
1 2 3 4
Quarter

Figure 2.21. Mean duration of visits by Anopheles gambiae females to the
upwind screen of a wind tunnel, broken out per quarter of the 30 min
observation period. This test was conducted in a wind tunnel using

either fully untreated or fully ITN-treated upwind screens.

56

screen of the wind tunnel did not change significantly for either Olyset or Permanet

across

all four quarters (Figure 2.22).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

:2 10
£1: Bl +Control
3‘ E —o—Olyset
: 3 6-
e : +Permanet
: g .
as 4-
,: 3 2 _ NS
C r
g 0 T\, “ETA . ‘3
1 2 3 4
Quarter
2 E 420
O c: 360 ~
g e
_c_> g 300 —
'5 ca 240 —
3 E 180 ~
g g 120 - NS
5 g 60 - a;
U 0 r m 4 r ——""‘
1 2 3 4
Quarter

Figure 2.22. Mean total frequencies of arrival by Anopheles gambiae
females on the down-wind screen of a wind tunnel, broken out per quarter
of the 30 min observation period. This test was conducted in a wind tunnel

using either fully untreated or fully ITN-treated upwind screens

57

The sequential organization of behaviors expressed by An. gambiae in this wind
tunnel test was analyzed as transitional matrices. Three 10 x 10 transition matrices were
constructed for behaviors of An. gambiae in the vicinity of untreated and fully-Pennanet
and Olyset screens, respectively (Table 2.2, 2.4, and 2.6 ). Three other 7 x 7 transition
matrices were constructed for positions in the wind tunnel including the screens upwind

(Table 2.3, 2.5, and 2.7).

Table 2.2. A 10 x 10 transition matrix of behavioral responses of An. gambiae when

untreated netting covered the upwind end of a wind tunnel.

 

 

 

Behavior flying walk sitting k- n- 1- tap- Jerk- Frrati l-bp- w-sp- SUM
down flying ping ing o-m ping read

flying - 0 2L4 0 1_3_5 1 0 0 _7_3 0 463
walk 1_7 - 5_O 0 86 2_26 0 0 _1_7 0 3%
sitting _2§_2 Q7 - 0 £5 8_5 3 1 _2_1_7 1 1271
k-down 0 O 0 - 0 0 0 O 0 0 0
n-tapping 110 Q 6_53 0 - 5 0 0 14 0 802
l-tapping 26 £3 a 0 g - 0 0 32 0 320
jerking 2 0 I0 0 0 0 - 0 1 0 3
erratic-m 0 0 0 0 1 0 O - O 0 1
hopping 61 3 a 0 7 3 0 0 - 0 362
w-spread 0 0 0 0 1 0 O 0 0 - 1
SUM 458 394 1298 0 73) 320 3 1 354 1 3619

 

Bolded and underlined values are significantly higher than expected by chance, while underlined
values are lower than expected by chance. Bolded only values are moderately higher than expected
by chance, after Bonferroni correction. Values in normal font are not significant. Italized values had
expected value < 5 and n0 chi square test was calculated. K-down = knockdown; n-flying = net flying;
l-tapping = labella tapping; erratic-m = erratic movements; w-spread = wings spread.

58

Table 2.3. A 7 x 7 transition matrix of positions of An. gambiae behaviors when

untreated netting covered the upwind end of a wind tunnel.

 

 

 

Position t-screen wirescreen n-side n-roof plexiglas floor intersection SUM
u-treated - 1 0 5 2 5 _4_9_ 36 1 07
d—mesh 1 0 - ' 4 0 9 16 27 66
n-side 3 3 - 0 1 7 6 20
n-roof 2 0 0 - 2 1 0 5
plexiglas _§ 1 6 5 0 - 0 41 1 17
floor 2 1 1 0 0 0 - 4 17
refirge 33 24 4 O 3 fl - 1 1 1
SUM 105 64 18 2 20 120 114 443

 

Bolded and underlined values are significantly higher than expected by chance, while
underlined values are lower than expected by chance. Bolded only values are moderately
higher than expected by chance. after Bonferroni correction. Values in normal font are not
significant. Italized values had expected value < 5 and no chi square test was calculated. u-
screen = untreated screen; n-side = netted side; n-roof = netted roof

Table 2.4. A 10 x 10 transition matrix of behavioral responses of An. gambiae

when Permanet covered the upwind end of a wind tunnel.

 

 

 

Behavior flying walk sitting k- n- l- tap- Jerk- errati- hop- w-sp- SUM
down flying pirlg ing c-m ping read
flying - 1 2_71 6 1 06 1 0 L 33 1 454
walk 7 - 18 0 20 42 2 1 1 7 0 1 07
sitting _2_9_2_ 1Q - 8 291 29 gr; 1 3 113 1 5 926
k-down 1 0 1 - 0 0 4 26 0 5 37
n-tapping 65 6 _3_§3 1 - 0 0 1 0 14 2 460
l-tapping 12 25 22 0 6 - 0 1 1 0 0 76
jerking 9 4 1 6 4 1 - 4 4 4 37
erratic-m 22 O 25 22 1 3 0 1 - 7 1 1 101
hopping 24 1 m 3 5 3 0 1 - 1 266
w—spreed 8 0 5 6 0 O 4 1 2 0 - 35
SUM 440 1 07 933 52 451 76 41 1 03 257 39 2499

 

Bolded and underlined values are significantly higher than expected by chance, while underlined
values are lower than expected by chance. Bolded only values are moderately higher than
expected by chance, after Bonferroni correction. Values in normal font are not significant. Italized
values had expected value < 5 and no Chi square test was calculated. k-down = knockdown; n-
flying = net flying; l-tapping = labella tapping; erratic-m = erratic movements; w-spread = wings
spread.

59

Table 2.5. A 7 x 7 transition matrix of positional placement of An. gambiae when

Permanet covered the upwind end of a wind tunnel.

 

 

 

Position t-screen wirescreen n-side n-roof Jlexiglas floor intersection SUM
treated - 6 9 1 4 _33 37 96
d-mesh § - 1 1 2 1 9 §_0_ 58
n-side 8 0 - 0 1 6 2 1 7
n-roof 3 0 2 - 1 1 1 8
plexiglas 28 1 9 3 O - O Q 89
floor 1 6 1 1 0 - 5 14
refiige 4; 26 3 1 3 37 - 1 1 2
SUM 87 57 19 4 1 1 102 114 394

 

Bolded and underlined values are significantly higher than expected by chance, while underlined
values are lower than expected by chance. Bolded only values are moderately higher than
expected by chance, after Bonferroni correction. Values in normal font are not significant.
Italized values had expected value < 5 and no chi square test was calculated. u-screen =
untreated screen; n-side = netted side; n-roof = netted roof

Table 2.6. A 10 x 10 transition matrix of behavioral responses of An. gambiae

when Olyset covered the upwind end of a wind tunnel.

 

 

 

Behavior flying walk sitting k- n- 1- tap- Jerk- Errati Hop- w-sp- SUM
down flyirg ping ing -c-m ping read
flying - 0 £6 4 fl 0 0 m E 16 739
walk 25 - 49 0 12 50 1 4 1 5 0 1 56
sitting SE 99 - 16 £5; 52 69 26 3g 33 1489
k-down 2 0 0 - 1 O 8 40 5 9 65
n-tapping 130 3 _4_2_7 2 - 1 2 6 1 5 9 595
l-tapping 17 40 20 0 12 - 4 0 14 0 107
jerking 30 10 2 6 12 2 - 63 28 6 159
erratic-m 46 0 103 46 23 O 14 - 12 9 253
hopping 49 3 §_8_3_ 1 1 3 1 2 21 - 2 595
w-spread 3 0 2 5 1 0 5 1 0 4 - 30
SUM 715 155 1502 80 595 106 105 277 549 84 4168

 

Bolded and underlined values are significantly higher than expected by chance, while underlined
values are lower than expected by chance. Bolded only values are moderately higher than
expected by chance, after Bonferroni correction. Values in normal font are not significant. Italized
values had expected value < 5 and no Chi square test was calculated. K-down = knockdown; n-
flying = net flying; l-tapping == labella tapping; erratic-m = erratic movements; w-spread = wings
spread.

60

Table 2.7. A 7 x 7 transition matrix of positions of An. gambiae behaviors when

untreated netting covered the upwind end of a wind tunnel.

 

 

 

Position t-screen wirescreen n-side n-roof plexiglas floor intersection SUM
treated - 17 12 9 1 0 fl §§ 1 90
d-mesh 16 - 4 2 3 26 31 82
n-side 16 6 - 2 2 9 4 39
n-roof 9 1 4 - 0 2 2 1 8
plexiglas 6 7 1 1 - 8 4 27
floor IQ 21 1 3 4 9 ~ fl 1 87
refuge 6;; 1 9 6 1 0 14 .6_1 - 1 75
SUM 182 71 40 28 38 180 179 718

 

Bolded and underlined values are significantly higher than expected by chance, while underlined
values are lower than expected by chance. Bolded only values are moderately higher than expected
by chance, after Bonferroni correction. Values in normal font are not significant. Italized values had
expected value < 5 and no chi square test was calculated. t-screen = treated screen; n-side = netted
side; n-roof = netted roof

Based on a 2 x 2 contingency table analysis, nine before/after transition
frequencies of behaviors on untreated screen were highly significantly elevated over
values expected by chance and two before/after transitions were moderately significantly
higher than expected by chance (Table 2.2). Transitions in position from Plexiglas to the
untreated screen, from the untreated screen to the floor, and from intersections to the
floor were pronounced (Table 2.3). This indicated that the relevant transitions of these
frequencies of behaviors required further organization. A kinematic graph (Figure 2.23)
was constructed revealing four major sequences and two minor sequences of An. gambiae
female behavior on an untreated screen in a wind tunnel: a) flying and sitting sequence
occurred mainly between the untreated screen and the floor, intersections and the floor,
and Plexiglas and the floor; b) a net-flying (tapping) and sitting sequence was confined

on the upwind untreated screen and it occurred at the highest frequency; 0) a walking and

61

labella tapping sequence was most evident after sitting on the untreated screen; (1) a
hopping and sitting sequence was evident on the screen and on the floor; e) a one-way
labella tapping to flying sequence was just significant on the untreated screen; and f) a
one-way hopping to flying sequence was intermittently mingled with other sequences.
Other sequences represented in Figure 2.23 were insignificant.

When the wind tunnel contained a Permanet screen, nine before/after transition
frequencies of behaviors were significantly higher than expected by chance, and two
before/after transitions were moderately significantly higher than expected by chance
(Table 2.4); they underwent further organizational analysis. Transitions in position from
the Permanet screen to the floor, from the downwind, metal screen to intersections, and
from intersections to the Permanet screen were significantly higher than expected by
chance (Table 2.4), and so were organized as explained below.

A kinematic graph (Figure 2.23B) was constructed revealing four major
sequences and one minor sequence of An. gambiae female behavior on the Permanet
screen in a wind tunnel: a) a flying and sitting sequence occurred mainly between the
Permanet screen and downwind metal screen, Permanet screen and the floor and
intersections, and Permanet, b) a net-flying (tapping) and sitting sequence was confined
on the Permanet screen and occurred at lesser frequency than on an untreated net, 0) a
sitting jerking sequence was most evident on the Permanet screen, (1) a hopping sitting
sequence was evident on the Permanet screen and on the floor, e) a net flying (tapping)
flying sequence was modestly significant both ways on the Permanet screen, and t) a

one-way hopping to flying sequence was intermittent with other sequences.

62

The Olyset screen yielded nine before/after transition frequencies of behaviors
that were significantly higher than expected and three before/after transitions, which were
moderately significantly higher than expected by chance (Table 2.6). Transitions in
position from the Olyset screen to the floor, from the Olyset screen to intersections and
back to the Olyset screen , and from the floor to the Olyset screen, were significantly
higher than expected (Table 2.7) and were therefore were organized into a kinematic
graph for determination of sequences of behavior.

A kinematic graph (Figure 2.23C) was constructed revealing five major sequences
and gambiae female behavior in a wind tunnel presenting an Olyset screen: a) a flying
and sitting sequence occurred between the Olyset screen and the floor, Olyset screen and
intersections, b) a net-flying (tapping) and sitting sequence was confined on the Olyset
screen and occurred at lesser frequency than on the untreated net but more than on the
Permanet screen, c) a sitting jerking sequence was most evident on the Olyset screen, d) a
hopping and sitting sequence was evident on the Olyset screen and on the floor, e) a net
flying (tapping) and flying sequence was slightly significantly both ways on the Olyset
screen, i) a one-way flying to erratic movements sequence was prominent, g) a one-way
sitting and labella tapping and jerking sequence occurred mainly on the Olyset screen,

and h) a one-way sitting to walking sequence was significant on the Olyset screen.

63

21

 

 

 

 

 

 

   
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  
   

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NET TAPPlNG 652. Sl'lTlNG 50 WALKING I.162 waurme
403 327
005 ‘35 1328 707'" 2.2.6,,
135 HOPPING 54
110
373" [5:42 26
FLYING
NET TAPPING ...362 SITnNG _.70 WALKING 1?. mammals-re
525 ..__ 110 7g
‘57 1065 18 F25
172{ ’22;
HOPPING JERKING
200 77
65
33] 24
,
FLYING ERRATIC now
‘32—- ‘22
01 L1 261 1,22
WNGSPREAD KNOCKDOWN
59 56
C
NET TAPPING _5427 SFFHNG 99 _, muons 3‘3. maumapue
597 ..__. 160 ._ 100
.385 1532 49 40
396 L483 59
HOPPING 103 JERKING
534 413 416 26
130 50] 149 14] [63
31 L16 40 1 L45
WHGSPREAO KNOCIGJOWN
34 80

Figure 2.23. Kinematic graphs of behavioral sequences of Anopheles gambiae in a

wind tunnel with upwind screens configured immediately downwind of host cues.

Arrows represent the transitions between behaviors; numbers next to the arrows refer to
frequencies of transitions. Boxes represent behavioral states and numbers inside the boxes refer
to frequencies of behavior. A = untreated screen: B = Permanet screen: C = Olyset screen.

64

T est k Untreated gage vs half-treateg_D_e_l_tg_zm ethrin1 my! Perrnethrin screens

There was no evidence that An. gambiae females detected and avoided
Deltamethrin- and Perrnethrin-treated sections of half-treated cages. The profiles for the
treated and untreated halves of the screen in the wind tunnel were statistically identical
for all screens containing insecticide (Figure 2.24). This was true throughout all four
quarters of the 30-min tests. Lack of significant difference in frequency of arrival on
either side of the half-treated screens during all four quarters was previously reported for
half-treated cages (Figure 2.11 and 2.12). The inability of An. gambiae to differentiate
between the sides of the half-treated wind tunnel treated with WHO recommended
dosages of Deltamethrin and Perrnethrin argues strongly against repellency by these
insecticides.

Duration of stay after arrival per half of upwind screens was significantly lower
for the Perrnethrin-treated side relative to a fully-untreated screen in the wind tunnel.
However, An. gambiae did not discriminate between the untreated half of Perrnethrin
treated screens (Figure 2.25). The significantly shorter stay on Perrnethrin half-screens
relative to the control screen argues for some perception of chemical presence, possibly

via some irritant effect.

65

A

 

2.5
QUARTER 1

1.5‘

0.5‘

 

NS; F=0.36

 

o Untreated
I Treated

 

 

 

I0

0

 

 

Control

C

Deltamethrin Perrnethrin

Screen in Wind-tunnel

 

° QUARTER 3

1.5

0.

Mean frequency of Arrival per Half of Upwind Screen

0.5

 

NS; F=O.13

I0

0-

 

 

Control

Deltamethrin Perrnethrin

Screen in Wind-tunnel

Mean frequency of Arrival per Half of Upwind Screen

 

 

 

 

 

2° QUARTER 2 o
2‘ _
1.5 3
1l .
0 NS; F=O.60
0.5%
0 fl 1
Control Deltamethrin Permethr'n
Screen in and-tunnel
D
2.5
QUARTER 4
2‘ NS; F=O.42
1.5.
1 .
8 :
0.5l
0 fi r

 

 

 

Control Deltamethrin Perrnethrin
Screen in Wind-tunnel

Figure 2.24. Mean total frequency of arrivals by Anopheles gambiae on

upwind screen in a wind tunnel having screens either fully untreated or half-

treated with deltamethrin or permethrin, as broken out by quarters of the 30-

min observational period.

66

>

 

N
(h
o

s

_g
(h
0

§

0!
O

 

 

 

 

 

 

 

O

QUARTER 1 ° Untreated
3 a 'Treated
a,b
2b a,b
I
ah
I
Control Deltamethrin Perrnethrin

Screen in Wind-tunnel

 

Mean Duration per Half of Upwind Screen
§ § '§ § §

0|
0

 

 

 

o

a a QUARTER 3
Ob ob
lb I b
Control Deltamethrin Perrnethrin

Screen in Wind-tunnel

(D

N
0'1
O

0

Mean Duration per Half of Upwind Screen

0

 

's’

_s
or
o

..a
O
O

(’1
O

 

 

 

a
5 QUARTER 2
b b
I 0
0b b
I
Control Deltamethrin Perrnethrin

Screen in Wind-tunnel

 

M
O
O

_a
U!
o

5

0'1
0

 

 

 

a 8 QUARTER 4
Ob b
b o
' I
b
Control Deltamethrin Perrnethrin

Screen in Wind-tunnel

Figure 2.25. Mean total durations of visits by Anopheles gambiae on

upwind screen in a wind tunnel having screens either fully untreated or

half-treated with deltamethrin or permethrin, as broken out by quarters of

the 30-min observational period.

67

DISCUSSION
Main pre-knockdown effects of I T Ns on An. gambiae behaviors

Whether in fully-treated or half-treated static cages, or in a wind tunnel with a
fully—treated or half-treated upwind screen, ITNs profoundly affected the behaviors of An.
gambiae females within 15 min and killed many of them by 30 min. This was expected
based upon the large body of previous toxicological data (Bondareva et al., 1986;
Aldridge, 1990; Kolaczinski and Curtis, 2000; Mathenge et al., 2001; Hougard et al.,
2003). However, the main interest of the current study was on changes in behavior
occurring especially in the first but also in the second quarters of the 30 min observations,
before first knock-down. If a mosquito were to adjust its behaviors so as to escape the
lethal effects of ITNs, such adjustments would need to come early rather than late in the
sequence.

Seemingly, the most effective behavioral adjustment An. gambiae females could
make to avoid the lethal effects of ITNs would be to avoid alighting upon treated surfaces
or to depart quickly when contacting insecticide. However, no such evidence was found.
Foregoing data indicate that unexposed females alighted on treated surfaces as soon as
and as frequently on treated as on untreated surfaces (Figure 2.12 and 2.24). They
reduced their sitting time (Figure 2.18D) and females having some exposure to the ITNs
likewise did not avoid landing upon treated surfaces (Figure 2.13 and 2.25). Indeed, no
evidence was obtained that An. gambiae, at any time, was capable of discriminating
between treated vs. untreated surfaces. This inability was particularly telling in the wind
tunnel tests where the frequency of flying upwind and alighting on treated netting

equaled that for untreated netting (Figure 2.24).

68

Once mosquitoes had alighted upon ITN-treated surfaces and absorbed some
insecticide, behavioral effects were noted well before first knockdown. Chief among
them was an elevation in the frequency of flight initiation (Figures 2.5A, 2.11A, and
2.18A). Notably, mosquitoes exposed to insecticides in both static cages and the wind
tunnel and destined to spend a large proportion of the total 30 min in knockdown,
managed to accumulate as much flying time as did the comparable controls in their full
30 min with no knockdown (Figures 2.3, 2.9, and 2.16). Although, this locomotor
stimulant effect (Figure 1.6), did not extend to walking (Figures 2.5E, 2.11E, and 2.18E),
excitation was evidenced in elevated labella tapping (Figure 2.7).

These pre-knockdown behavioral effects of ITNs were subtle rather than
dramatic. This is reflected in the high similarities rather than dramatic differences
between the kinematic graphs for untreated controls (Figure 2.24A) vs. mosquitoes
exposed to Olyset or Permanet ITNs (Figure 2243 and C). Rather than revolutionize the
structure of behaviors, ITN exposure only shifted the fi'equency of transitions among pre-
knockdown behaviors and then added new behaviors associated with toxicosis.

Any evidence for avoidance?

No evidence at all was found for avoidance of ITNs in the static cage or wind
tunnel apparati, all of which were small and impermeable to mosquitoes. Had the static
cages and wind tunnel contained ports for escape, it is possible that the exit rate relative
to that for control mosquitoes could have been elevated due to the locomotor stimulant
effect of the ITNs. Such an effect would be similar to raising the temperature of a volume
of gas molecules contained in a leaky vessel; a rise in temperature would increase the rate

of efflux over that at a lower temperature. However, in a closed container, the rate of

69

interaction of gas molecules with all walls of the container would rise with temperature.
But because escape from a closed container would be blocked, the only overall response
would be an increase in pressure.

This line of reasoning suggests that expression of avoidance by an indirect
mechanism like the action of a locomotor stimulant could be context dependent. The
current experiments only addressed the situation where behavioral arenas were closed. In
this context, there was no evidence of avoidance mediated by either taxes (repellency) or
kineses (locomotor stimulant or deterrent) (Figure 1.6).

Mode of action of I T Ns

In the absence of any behavioral avoidance in static cages and the wind tunnel, all
evidence supports lethality as the only mode-of-action of ITNs in these given arenas.
Fit of current data with previous reports

The current data conflict with a number of previous reports suggesting that ITNs
function mainly by causing mosquitoes to avoid entering houses equipped with ITNs
(Lines et al., 1987; Snow et al., 1987; Lindsay et al., 1991; Chandre et al., 2000; Takken
2002; Gimnig et al., 2003) or by raising exit rate (exophily) once mosquitoes encountered
ITNs inside houses (Lockwood et al., 1984; Threlkeld 1985; Rutledge et al., 1999;
Chareonviriyaphap et al., 2001; Mathenge et al., 2001). Several reasons can be suggested
for the differences in outcomes of the current study relative to those reported previously:
1) the pyrethroids used here were highly pure (not accompanied by solvents or any other
carriers) and of larger rather than smaller molecular weight, 2) the long-lasting net
formulations used here, where insecticide is deeply imbedded in the fabrics by factory

processes, may reduce the surface concentrations and rates of evaporation of insecticide

7O

from ITNs as compared to nets treated only by dipping in insecticide solution, 3)
detection of avoidance might require presence of sizable openings in the apparatus
through which mosquitoes can readily exit, and 4) all mosquitoes in the current study
were naive and never previously exposed to any insecticide; perhaps pronounced
avoidance requires learning.

Finally, we note that some other studies report results similar to those reported
here. For example, Cuzin-Ouattara et al. (1999) and Maxwell et al. (1999) reported
repeated contact of many mosquitoes with ITNs resulting in mass killing as expressed by
a significant reduction in mosquito densities in treated areas. Miller et al. (1991)
reported different killing abilities by different ITNs and Guillet et al. (2001) showed that
An. gambiae that entered huts with nets treated with only deltamethrin (Permanet) or
bifenthrin 50mg a.i./m2 gave high mortality-rates (81-86%). The above observations
point out that interpretation of the effects of ITNs will likely require careful
documentation of both the mosquitoes being used in a study, as well as the spatial details
of the apparatus. Answering the central question of what is the mode-of-action of ITNs in
actual house settings will require that the tests be done under conditions exactly reflecting
the houses in a given area where ITNs are deployed. The next chapter of this dissertation
attempts to reflect housing conditions of the Luo people of Western Kenya near the city

of Kisumu.

71

CHAPTER 3
Quantifying Redistribution and Survivorship of Anopheles gambiae in a Multi-
Room Choice Arena Presenting Human Sleepers under Insecticide-Treated vs.

Untreated Bed-Nets

INTRODUCTION

Insecticide-treated bed nets (ITNs) have proven efficacious in reducing malaria
transmission, and in lessening morbidity and mortality in children in Sub-Saharan Africa
(D’Alessandro et al. 1995; Binka et al. 1996; Nevill et al., 1996; Habluetzel et al., 1997;
Lengeler et al., 2000; Schellenberg et al., 2001; Gimnig et al., 2003a, Phillips-Howard et
al., 2003b). Entomological evaluations have documented population suppression of
Anopheles gambiae Giles and An. funestus in the intervention areas (Gimni g et al.,
2003a), and a positive community-wide improvement in health of occupants in nearby
houses without ITNs (Hawley et al. 2003b). ITNs have been adopted by the World
Health Organization (WHO) as efficient tools for malaria vector control and are now
utilized on an increasingly large scale in the Roll Back Malaria initiative (WHO, 1993;
RBMP, 2001). In the face of projected widespread and long-term distribution of ITNs,
attention is shifting from immediate efficacy to sustainability, with emphasis on new and
long lasting net formulations.

Surprisingly, knowledge is limited as to how IT Ns affect mosquitoes, which is
one of the key issues in addressing their long-term impact. Yet understanding how to
manage ITNs for sustainability requires understanding their mode of action, including the

behavioral interactions between mosquitoes and ITNs.

72

Two main explanations have been postulated for ITN mode-of-action. First is a
killing effect, whereby the density and mean age of local populations of vectors is
reduced and transmission drops accordingly (Magesa et al., 1991; Guillet et a1, 2001).
Mosquitoes visiting ITNs while host seeking were envisioned to absorb sufficient
dosages of toxicant to cause knock-down and death, even after they may have flown off
the net and displaced to other locations inside or outside a house. This outcome is
revealed in “cone tests” (Kolaczinski and Curtis, 2000; Hougard et al., 2003), where
mosquitoes are confined in small plastic cones clipped onto ITNs. However, in actual
African houses, finding dead mosquitoes on the dirt floors is difficult, given the
considerable visual interference and presence of scavengers such as ants. Thus, definitive
evidence for direct lethality of ITNs under house-use has been weak, and its
demonstration required forced contact assays, that may not reflect what happens
normally.

Other researchers contend that the primary effect of ITNs is to cause avoidance of
the treated net, typically measured by demonstration of contact or non-contact “excito-
repellency” and/or “contact irritancy”, as reflected by reduced entrance into houses or
rooms in which ITNs are hung, and correspondingly increased exophily or exit rates from
these domiciles (Lines et al., 1987; Lindsay et al., 1991; Mathenge et al., 2001). The
terminology used to describe the consequences of exposure to ITNs by direct contact or
by proximity but non-contact is diverse and sometimes confusing. Vocabulary is often
influenced by the device used to conduct a test, and has a long historical context
extending to evaluation of indoor spraying of insecticides for malaria control. The term

“avoidance” is a reasonable word used to capture this non-lethal effect. The various

73

apparatus for these studies have included tunnel tests and hut studies, both attempting to
model small and large-scale mosquito behaviors in response to treated nets in the
presence of host cues. For hut studies, the protocols have varied between East and West
Africa. In West Africa, entrances are narrow slits located at window-height; exit traps are
conical and placed within doorframes. In East Africa, entrances are the house eaves; exit
traps are placed in front of window frames. In both settings, test designs have generally
used multiple houses arranged in rows, and have relied on visitations by wild mosquitoes,
usually in locations where adult production is heavy, such as fi'om rice fields. Variation
can be imparted by: heterogeneous spatial distribution of mosquitoes, shifts in breeding
sites during a test, subtle differences between huts, as well as but structural designs.

Whether avoidance, lethality, or a combination of the two is most desirable for
optimal and sustained ITN function is still unclear. Both may operate in implementation
programs, but the degree of one effect or the other could be a function of formulation of
insecticide on the net and time. Additional studies are needed that focus on the newer,
long-lasting ITNs (Long Lasting Impregnated Nets or LLINs) that are and will be
increasingly deployed for the next decades.

This study reports on tests using a new, house-like apparatus, where conditions
are well-defmed. Mosquitoes are released into a common courtyard connected to four
sleeping rooms only by cave-like openings at 2 m height. This type of design permits
manipulation of number, age, and strain of mosquitoes. Also, care can be taken that
rooms are equivalent. The study addressed interactions of An. gambiae with Permanet,

Olyset, and NetProtect ITNs. The objective was to determine whether the data supported

74

avoidance vs. lethality of ITNs deployed in a house-like apparatus with large eve-
openings typical of Luo housing in Western Kenya.
MATERIALS AND METHODS
Apparatus
A four-choice apparatus permitting overnight sleepers in a secure arrangement
was constructed at the Center for Vector Biology and Control Research (CVBCR) in
Kisian, Kenya; it is a field station of the Kenya Medical Research Institute, a national

biomedical research organization. It was named the Mosquito Hotel (Fig. 3.1). To

 

Figure 3.1. Photograph of the Mosquito Hotel on site, showing the main

entrance to the central courtyaid, and the exterior of two of four sleeping

rooms.

75

preclude ants and other predators/scavengers, the structure was built on a platform,
supported by 1 m posts resting in 10L metal cooking pots each containing ca. 7 L of used
automobile engine-oil.

The floor plan is given in Fig. 3.2. Walls and rafters were 5 cm x 10 cm wood
construction. Sleeping rooms were roofed by corrugated iron sheets. The exteriors of
sleeping rooms were plywood, painted white. The interior walls of the sleeping rooms
were covered with white, 25 mm thick Styrofoam sheets for insulation. Further, the
insulation and plywood flooring were tightly lined with a removable, white polyester

textile (flexibly woven, 7O g/mz).

 

 

 

 

 

 

Sleeping
Room
1 D 1
l l
4+
a— __
24", Sleeping D36,“ Courtyard D Sleeping
' Room __‘ __ Room
V

 

 

 

 

 

 

 

 

‘-—2.4m—7 y

 

Sleeping
Room

 

 

 

Figure 3.2. Floor plan of the Mosquito Hotel showing the configuration of

the central courtyard and sleeping rooms. D = door.

76

Doors leading from the courtyard to each sleeping room measured 1.8 x 0.9 m.
They were fashioned fiom textile split with a vertical zipper. Beds were 1.8 x 0.9 m with
a foam mattress set 60 cm above the floor on a wooden fi'ame. Rectangular nets were
hung by strings anchored at the four comers where the roof joined side-walls. During
tests, nets were tightly tucked under the mattress. Mosquitoes were able to access spaces
under the beds. Above each door was a 20 cm high x 110 cm wide “simulated eave”
opening for mosquitoes to enter or leave a sleeping room (Figure 3.3.a.). This opening
was fitted with the restrictor apparatus shown in (Figure 3.3.b). The top of the restrictor
consisted of netting attached to a bar that could be moved to adjust the opening from 3 to
20 cm. The screening allowed for similar levels of air exchange at any opening size.

The floor of the 3.6 x 3.6 m courtyard was also textile-covered; its ceiling was
sealed with nylon mosquito netting (mesh 50 g/mz). A 6 x 6 m green canvas completely
covered the courtyard and extended over the comers of the sleeping rooms (Figure 3.1).
Air exchange occurred between the canvas courtyard roof and netted ceiling. The space
between sides of sleeping rooms and courtyard comers was covered with untreated
mosquito netting over fabric. Between tests the fabric was unzipped to enhance
ventilation of the courtyard. Night-time light intensity in sleeping rooms was 0.13 - 0.24
lux, while that in the courtyard was 0.91 — 1.30 lux; it was provided by security lamps in
the vicinity of the Mosquito Hotel.

Three brands of long-lasting insecticidal net (Permanet, Olyset and Netprotect)
and a net treated with alphacypermethrin (Fendona) were used in this study. Permanet,
manufactured by the commercial textiles firm Vestergaard-Frandsen, is made of polyester

fibers which contain ca. 55 mg deltamethrin per square meter of fabric, and is designed

77

WWMMQ-um0)~<o.-§o

 

Figure 3.3.a. Simulated eave opening above the entrance of a sleeping room

Screening

    

Figure 3.3.b. Entrance restrictor with screening and movable bar for

adjustment of eave opening to 3. 10 or 20cm. The smaller of the two openings

projected into the sleeping room.

so that the insecticide remains effective for 3-5 years. Netprotect has similar composition
to Permanet; it has a polypropylene fiber. It was provided by the Intelligent Insect

Control Company. The Olyset net marketed by the Sumitomo Chemical Company,

78

consists of ca. 1 g of permethrin incorporated per square meter of polyethylene fiber.
Fendona is a recent long-lasting insecticidal net from the BASF Corporation; it uses
alphacypermethrin at 25 mg per m2 of fabric. Both Permanet and Olyset have been
approved by the World Health Organization Pesticide Evaluation Scheme (WHOPES) for
use in malaiious areas. Netprotect and F endona are still undergoing evaluations and
efficacy testing.
Test protocol

Nets to be tested were hung outdoors in the shade for two d prior to deployment in
sleeping rooms. This permitted release of any immediately volatizing active ingredient
built up on surfaces while in the packaging. Preparations for a nightly test began at ca.
15:00 h when beds were prepared and nets were draped over them. Using hand-held
Sprayers, the floor and wall fabric linings of sleeping rooms were wetted with well water,
along with the courtyard floor. Each room received 3 L and the courtyard floor received 6
L. Typically, two hundred 2-3 day-old mosquitoes from the Kisumu laboratory strain
reared at the CVBCR of KEMRI, fed only 10% sugar water, were carried to the test site
in 11cm diam. x 15 cm high paper cups (50 per cup) with screened tops. Mosquitoes were
placed near the wet courtyard floor and allowed to acclimate. At about 21 :00 h four
sleepers entered the courtyard, opened the paper cups, and retired to their assigned beds
for the night, making sure that their door was zipped and the eve-entrance cover was
open.

At 6:30 h sleepers arose and closed the entrance apparatus; zipped all doors, and
departed. Between 7:30 —8:30 h, 2-4 technicians arrived and carefully collected and

counted all dead and alive mosquitoes from the courtyard and each sleeping room.

79

Search for mosquitoes was intensive and involved tiu'ning over and searching beneath the
beds. Dead and alive mosquitoes were. counted. Live mosquitoes were transferred to
previously unused, 11 cm diam. x 15 cm holding cups and given access to 10% sugar
solution and held in ca 75% relative humidity and subdued light in the laboratory. After
24 h, a final count was taken of the numbers of live vs. dead mosquitoes. Thus, for any
given test, the number of mosquitoes occupying each room was recorded, as was the
number alive and dead the morning after the experiment. The number found dead upon
the first morning after release will sometimes be referred to as “immediate mortality”,
and the number dead after 24 h as “delayed mortality”.
Room equivalence tests and effect of opening height

To test for equivalent performance of rooms, untreated nets were hung in all four
rooms. Four sleepers were randomly assigned to rooms the first of four nights. Sleepers
then rotated clockwise on successive nights until all persons had slept in all rooms. This
procedure was done with entrance openings set at 3, 10, and 20 cm.
Experimental design for testing I T Ns

In most experiments reported here, two rooms were assigned control (no-
insecticide) nets, and two were assigned treated nets, except in the case where three
treated nets were used. Again, each of four sleepers rotated through all rooms. Data
recorded each test day were: dead and alive mosquitoes in courtyard and each room,
blood fed vs. non-blood fed, additional mortality after 1 d, sleeper identity, and net type
in the room. A temperature and relative humidity profile were recorded by HOBO data-

logger placed in each room and courtyard at a height of 1.5 m. After a round of testing,

80

beddings and fabric room linings were removed, soaked in quatemium detergent
(A.I.S.E., Brussels) for 1 d, and flushed in clean water, and dried in the sun before reuse.
Effects of An. gambiae strain and age

The Kisumu laboratory strain and F1 progeny of wild-caught An. gambiae were
reared at the CVBCR of KEMRI, and adults were fed only 10% sugar solution. The
above-described test protocol was followed. To test for their survival and equivalent
occupancy of rooms, untreated nets were hung in all four rooms. Four sleepers were
randomly assigned to rooms the first of four nights. Sleepers then rotated clockwise on
successive nights until all persons had slept in all rooms. A total of 12 replicates for each
strain was performed.

Experiments on the effect of mosquito age were similarly performed but used the
Kisumu laboratory strain only. The 16 day-old females were previously blood-fed, then
allowed to lay eggs; then they were kept hungry and maintained on 10% sugar solution
only for, 3-4 days before the test. Young mosquitoes were as described above.
Variations in site of release

To assess the extent to which mosquitoes visiting sleeping rooms were likely to
exit to the courtyard, four tests were done where 50 females were released into each
sleeping room and four where all 200 females were released into one room with no
releases into the courtyard. All other parts of the above protocol were followed and all
cloth linings of the sleeping rooms and courtyard were changed after using each of three

net types: Permanet, Olyset, and Netprotect.

81

Data analysis
Data were analyzed by AN OVA, implemented in SAS (2000). Data were

proportions derived from counts of mosquitoes, and were transformed by the arcsine of
the square root. For room occupancy, the denominator in the pr0portionate data was the
number released in the courtyard. For mortality, the denominator was the number alive
and dead in the room the morning after the trial. To account for variations among rooms
and sleepers, each room was considered a separate treatment in the analysis, even if the
same type of net was hung in that room, as sleepers were rotated. Means were separated

using Tukey's multiple comparisons test with a significance level of P < 0.05.

RESULTS

T emperature/RH profiles

Figure 3.4 shows selected plots of temperature and RH in the sleeping rooms over
the course of some experiments. They represent the highest and lowest average night-
time conditions in the test rooms and two mid temperatures. When the sleeper arrived in a
room about 21:00 h, the temperature rose slightly; when the sleeper exited, the
temperature fell slightly. After 6:30 h, when the rooms were evacuated and the sun
appeared, the temperature in the rooms rose steadily, while the RH fell. Figure 3.5
presents selected room temperatures and RH taken at varying times of year. Higher
temperatures and lower RH were evident in March than during subsequent months.
Average room temperature and RH fluctuations across the seasons were modest. Room

temperature fell to a minimum when sleepers exited the Mosquito Hotel, permitting some

82

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Roomt
100
5 9° E
on a
if so E
g g
I: 50 . .2 50
Time Time
100 Room3 100 Room‘
5 °°‘ g 90~
'0 on
t 80' ‘ E 00“
g 70‘ g 704
0
g 60‘ 3- 60*
,3 8
50 v v r v v v '— 50 r t . 1 f v
17:30 20:00 22:30 1:00 3:30 6:00 8:30 17:30 20:00 22:30 1:00 3:30 6:00 0:30
Time Time

Figure 3.4. Plots of temperature and RH in the sleeping rooms on a single
test day. Temperature is expressed in °F rather than °C for convenience in

co-plotting %RH with Temperature.

cooler outside air to enter the apparatus. Temperature and RH across the four sleeping

rooms therefore are shown to be equivalent.

83

 

 

Temperature VF) RH (1i)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Room in March, 2006 _ Room in April. 2008

100 100
a 90 E 90
no a 80
r: 9“ r:
0 A
3 70 g 70
'5 e
g- 50 g 301
E E
h SUI I I I I I I '2 50 I I I I I I

17:30 20:00 22:30 1:00 3:30 6:00 8:30 17:30 20:00 22:30 1:00 3:30 6:00 8:30

Time . Time
Room in May. 2008 100 Room in June, 2006

100
a 90 E 90.
a ‘1';
E 00 t 30.

O
O I—
h 7 ‘
a 70 g 0
9 s
0

g 50 I I I I I I I F 50 I I I I v i

17:30 20:00 22:30 1:00 3:30 5:00 9:30 17:30 20:00 22:30 1:00 3:30 5:00 0:30

Time Time

Figure 3.5. Plots of temperature and RH of sleeping rooms on different dates

and test months.

Room equivalence tests and effect of opening height

Tests carried out with untreated nets in rooms at a uniform cave opening height of
20 cm for all rooms, were statistically insignificant in mean percent occupancy at P<0.05
(Fig. 3.6). Room occupancy for certain sleepers was significantly higher than for other
sleepers. However, because sleepers were always rotated through all rooms, this effect

was inconsequential with respect to results and conclusions to follow.

84

 

3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

E a
G
5' 16%“ I; III +
E I
§ 8% -
it
c 4% 'l
B
0
E 0% . . .
Room 1 Room 2 Room 3 Room 4
Rooms with Untreated Bedneta

Figure 3.6. Mean percent Anopheles gambiae occupancy in rooms with
untreated nets for simulated eave opening height of 20 cm. Two hundred

mosquitoes were released on each of 12 nights of testing.

Mosquito occupancy in all four rooms with eave opening heights of 3 cm, 10 cm
and 20 cm are shown in Figure 3.7a. There was no significant effect of eave opening
height for entrance rate/occupancy of female An. gambiae 5.5. (P = 0.31). Similarly, the
mean percent mortality after 24 h was below 10% and not significantly different between
the three opening heights (Fig. 3.7b).

Effects of I T N type

There was no significant difference in occupancy rate of rooms when Permanet,
Olyset, and untreated nets were included in a replicated experiment with a central
courtyard release of mosquitoes (P = 0.4) (Figure 3.8a). However, mortality at 8 and 24 h
varied significantly with treatment (Figures 3.8b). Both 8 h and 24 h mortality rates were

significantly higher for Permanet than Olyset or controls (P < 0.0001).

85

E’s

 

 

 

 

 

 

 

 

 

 

 

E... '-
n ' ‘
§ . —+—— T
'5
§ 10%-
a
g 5%"
8
5 0% . . .
3cm 10cm 20cm
Eave Opening

Figure 3.7.a. Mean percent Anopheles gambiae occupancy in rooms with

untreated nets for simulated eave opening heights of 3, 10 and 20 cm. Two

hundred female mosquitoes were released in the central courtyard each of

12 nights of testing.

 

 

 

 

A
m
E
it“
2
fi"
‘5
84301
e
0.
52s a a
e an
Eur. J I . “I" - I l
360 10cm 20cm
EaveOpening

 

 

 

 

 

 

 

 

 

 

 

 

B

was
g

3%.
g a
i a:
§

“-1
e I a
t 1 1
re 2"" 1
e
E on . r

3 cm 10 cm 20 cm
Eave Opening

Figure 3.7.b. Mean percent Anopheles gambiae mortality in rooms with

untreated nets for eave opening heights of 3, 10 and 20cm. A is 8 h mortality, B

is 24 h mortality.

86

Mean Percent Mortality

Mean Percent Occupancy

 

 

 

 

 

 

 

 

 

 

 

 

 

 

26%
20% -
o i! 7'
15 /0 ‘ I E
10% -
5% .
0% , , - - _
Untreated 1 Untreated 2 Permanet Olyset

Rooms with Untreated vs. Treated Bednets
Figure 3.8.a. Mean occupancy of female Anopheles gambiae in sleeping
rooms with Permanet and Olyset vs. untreated bednets. Two hundred

mosquitoes were released in the central courtyard on each of 14 nights of

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

testing.
A B
100 100
so 0 3 so l
g b
60 o 50 . T
3 22.15:;
E 03$
40 b 3 ‘0 ‘ 5:22;;
5 c :x;
23323-3: a E-SjS-Egi
20 c g 204 d I §:§;§I§;§
:33 a x39
om—J=E“ ' I l 1, - an} :I O I i T ' 1 an:
Urureatedt UntreatedZPennanetOtyset Urrtreatedi umeatedZPermanetOlveet

Rooms with Untreated vs. Treated Bednets R001"! With Untreated V8- Treated 306001!

Figure 3.8.b. Eight h mortality (A) and 24 h mortality (B) of female Anopheles

gambiae in rooms with Permanet, Olyset and untreated nets.

87

The mean occupancy rate for rooms fitted with Permanet, Olyset or Fendona
ITNs vs. an untreated bednet was similar (P = 0.52) (Figure 3.9). Mortality at 8 h and 24
h was significantly higher for all three rooms with ITNs than for control (P < 0.0001)
Figure 3.10). However, the 24 h control mortality was unusually high in this test, possibly
due to mosquito handling holding problems. The overall mortality in rooms with treated

nets was over 95%.

26%

 

N
O
a?

10%* l

5%*

Mean Percent Occupancy

 

 

 

 

 

 

c
32

 

Untreated Permanet Olyset Fendona
Figure 3.9. Mean occupancy of female Anopheles gambiae in rooms with
Permanet, Olyset and Fendona vs. untreated bednets. 200 mosquitoes were

released in the central courtyard on each of 12 nights of testing.

Rooms with untreated nets, and those fitted with deltamethrin chessboard nets
treated with 30 x 30 cm patches of 55mg/m2 vs. entire 30mg/m2 nets, did not differ

significantly in room occupancy (P = 0.98) (Fig. 3.11). Figure 3.12 shows 8 and 24 h

88

>
El

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

em 2:

£100 a 5100‘

g so 13 so.

a a b

‘5 5° E 50‘

t .0. t .0.

O O

a. b o.

g 20 g 20-

0 o

I o . " . I o
Untreated Pennant»! Olyset rm Untreated
RoomswithUntreatedveJreatadBednete RoomewltnUntreatedveJreatadBamete

Figure 3.10. Eight h mean percent mortality (A) and 24 h mortality (B) of
female Aopheles gambiae in rooms with Permanet, Olyset, and Fendona lTNs

vs. an untreated bednet.

 

25%

re
‘3
§

15% i a
10%1 f

5%.

Mean Percent Occupancy

 

 

 

 

 

 

 

o
32

 

Untreated 1 I Untreated 2 ' Deltamethrin Deltamethrin
30 mg Uniform 65 mg Cheeeboard

Rooms with Untreated ve. Treated Bednete
Figure 3.11. Mean occupancy of female Anopheles gambiae in rooms with 30
mg deltamethrin and 55mg chessboard deltamethrin vs. untreated bednets. Two

hundred mosquitoes were released in the courtyard on each of 16 nights of

testing.

89

mortality, which were significantly elevated over those for the controls (P < 0.0001).

Mortality at 24 h was pronounced for both treated nets.

 

 

 

 

  

 

100 100
; not 3:“ so
§ § so
I 60 I
E E
3 40 a g 40
h
a a
C 201 g 20
8 a a 3
I 0 r-rI—I 1"“! . . . I 0

Untreated Untreated Deltamethrin Deltamethrin Untreated Untreated Deltamethrin Deltamethrin
1 2 30mg Uniform 55mg Chessboard 1 2 30mg Uniform 55mg Cheeeboard

Rooms with Untreated vs. Treated Bednets Rooms with Untreated vs. Treated Bednets

Figure 3.12. Mean 8 h percent mortality (A) and 24 h mortality (B) of female

Anopheles gambiae in rooms with 30 mg deltamethrin and 55mg chessboard

deltamethrin ITNs.

Effect of An. gambiae Strain and Age

No significant difference was found for occupancy (P = 0.49) or mortality rates (P
= 0.75) when P 1 progeny of wild-caught An. gambiae were tested against the Kisumu
laboratory strain of mosquitoes (Fig 3.13).

The mean occupancy rate for 16-day old females was slightly but significantly
higher than that for 4-day old females (P < 0.032). However, the mean mortality rate after

24 h was not significantly different for the two age groups (Figure 3.14).

90

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A :33
82
o a
53” g a
E 7 a I
3.20% a E 42‘ 1 l
3 r I
0 or .L J1 % .
9”” so; .
5 5 Kisumu VVIld F1s
g 1083' C
a.
2
g 5% :81, a a
0% »_ .
Kisumu VVlld F1s E 42‘
Strain of Mosquito :1 . ..
g or , ,
Kisumu Wild F1s

Figure 3.13. Meanoccupancy and mortality of the Kisumu laboratory strain vs.

F 1 progeny of wild-caught Anopheles gambiae females in rooms with untreated

bednets. (A) Mean percent occupancy, (B) 8 h mortality, (C) 24 h mortality. Two

hundred mosquitoes were released in the courtyard on eaCh of 16 nights.

CY

Mean Percent Occupa

25%

20% -
15% ~
10% .
5% -

0%

 

 

 

 

l-—lcr

 

I‘Flm

 

 

 

 

 

4-day olds

16-day olds
Age of Mosquito

Figure 3.14. Mean percent occupancy for 4 day vs. 16 day-old female Anopheles

gambiae. Two hundred mosquitoes were released in the courtyard on each of 12

nights.

91

Variation in release site

On average, more than 95% of the An. gambiae females released into sleeping
rooms were found there and no significant differences were found between Permanet and
Olyset ITNs vs. untreated bednets (Figure 3.15). Mortality at 8 h was significantly higher
for Permanet vs. control rooms (P < 0.0001) and for Olyset vs. control rooms (P <
0.0001), although mortality was higher with Permanet than Olyset. The 24 h mortality
was not significant different in either case fiom controls, possibly indicating problems

with handling or holding mosquitoes during this experiment (Figure 3.16).

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

A e
120 120
2 a E O O
m‘ r:- 13 a "’0' 1 I a
g 30 * g 80 .
5° ‘ so ‘
is g...
§ 20 l i 20 1 “-4
o , , o . . , . ....,.
Untreated Untreated Permanet Permanet Untreated Untreated Olyset Olyset
‘l 2 1 2 1 2 1 2
Rooms with Untreated vs. Treated m Rooms vain Untreated vs. Treated Bednets

Figure 3.15. Mean percent occupancy of female Anopheles gambiae in rooms with
Permanet vs. untreated nets (A) and with Olyset vs. untreated nets (B) when 50

mosquitoes were released into each room on each of 8 test nights.

92

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

 

 

100 100
g 30‘ a g 80‘
g =- g
: °°‘ : 601 a
E E l
§ ‘° ‘ g ‘° 1
5 b b 5 2° ' b b
'2’ o .Jil- i o rfi r+| . '
Urireeted Untreated Permanet Permanet Untreated Untreated 0'18“ 0'18“
1 2 1 2 1 2 1 2
Rooms with Untreated vs. Treated Bednets Rooms with Untreated ve. Treated Bednets
C D
100 . 100
I
g 80+ 2.? 80'
E " a
2° °°‘ a g 60] }.
a , a a
5 ‘° I l g 40 I r}
o ifs-:12: $222155:
: 20- 2 2°‘ “2215-52;
a 3 :«::-55« ”at
:3 0 T . a 0 . - " - ,
Urlreated Untreated Permanet Permanet Untreated Untreated Olyset Olyset
1 2 1 2 1 2 1 2
Rooms with Untreated ve. Treated Bednets Rooms with Untreated ve. Treated Bednets

Figure 3.16. Mean percent mortality of female Anopheles gambiae in rooms
with treated vs. untreated bednets. (A and B) = 8 h mortality; (C and D) = 24

h mortality, respectively.

There was no difference in occupancy, 8 h, or 24 hr mortality when eaves were

open or closed and mosquitoes were released inside the sleeping rooms (Fig. 3.17).

93

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Em
grow
80( '
120A ‘0‘ I
a g" ‘
am 3 ’°‘
:: 0
3. oo- 3 Open
. -~ C
*5 5° £120 a
g 404 100 l
8 w j
g 20. 30.
§ § 40‘
o c a),
open g 0
Open

 

Figure 3.17. Occupancy and mortality of Anopheles gambiae females in rooms

with eaves open or closed. A = mean percent occupancy, B = 8 h mortality, C =

24 h mortality.

When 200 females were released in one of the four sleeping rooms and their
appearance in the other sleeping rooms and the courtyard was assessed after 8 h, the
mean occupancy rate for the release room was 78%. A total of 7% of the emigrating
mosquitoes was equally distributed in the other sleeping rooms (Fig. 3.18), and the

remainder were recovered in the courtyard.

94

 

 

 

 

 

 

 

 

100%

E‘ 30% -

O

a.

3

3 80% -

O

3.3 40%

‘6

a.

s 20% 4

g , b b b

0% T m T m T 4%
Stocked Room Unatocked Unatocked Unetocked
Room 1 Room 2 Room 3

Figure 3.18. Distribution of Anopheles gambiae females in rooms with 20 cm
eave openings when all 200 mosquitoes were released in only 1 of 4 sleeping

rooms.

Netprotect

No significant differences were evident in occupancy rate of rooms by An.
gambiae females when Permanet and Netprotect ITNs were paired against untreated
bednets and mosquitoes were released into the courtyard (Figure 3.19). Mortality at 8 h
was significantly elevated for ITNs over the controls (P < 0.0007). At 24 h, mortality was
likewise significantly elevated (P < 0.028). Mortality rates for Permanet and Netprotect
were not significantly different from each other: However, 24 h mortality for Netprotect
tended to be greatest. For both Permanet and Netprotect, there was no evidence of

avoidance.

95

 

S
:2

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

 

 

 

 

 

 

 

 

 

 

 

8

5

0.

§ 410% 1 a a

O 3024‘ .

S ..

. 20% .

a.

5 10%. ,_ r.

O

E

0% T .
Untreated 1 Untreated 2 Pennenet
Rooms with Untreated ve. Treated Bednets
A B

E 120 E 120

grow g 100 I :e-

3 3 E

E 50* '5 50‘
° 5:52:25;

§ 401 g 40 b b

0. 5:35:33:

r: m 9 a a; 20 1 + ..

e b b i - g

i o . . , a o . ' _,

Uritreated1UntreatedzPermenet Hebrew "WWIItreabflPermnat
Rooms with Treated vs. Untreated Bednets Rooms with Treated vs. Untreated Bednets

Figure 3.19. Occupancy and mortality of Anopheles gambiae females in rooms
treated with Permanet and Netprotect vs. untreated nets. A = 8 h mortality, B =
24 h mortality. Two hundred mosquitoes were released in the courtyard on

each of 16 nights of testing.

When only Netprotect was deployed in two rooms vs. untreated bednets in the
remaining two rooms, the occupancy of An. gambiae females was slightly but not
significantly higher in rooms with Netprotect than in rooms with control (P = 0.13) (Fig.
3. 20). Both 8 h and 24 h mortality rates were highly significant relative to the control

bednets (P < 0.0001).

96

 

  

 

 

 

 

 

  

a. 25%
c
a 20% . a
3 a
3 15%« s
E
E 10%-
: 5’"
i or. . . ' -
Untreated 1 Untreated 2 Netprotect Netprotect
Rooms with Untreated vs. Treated Bednets
A
i E
E °° g °°
g to g 00
g 3
§ 40 E 40
a 8
5 2° : 20
5 o- 3

 

 

 

Untreatedttlnueetedzuetprotecttuetprotectz Untreatsdtumrestsdzuetprotscttiietprmz
Rooms with Untreated vs. Treated Bednets Rooms with Untreated vs. Treated Bednets

Figure 3.20. Occupancy and mortality of Anopheles gambiae females in rooms
treated with Netprotect vs. untreated nets. A = 8 h mortality, b = 24 h mortality.
Two hundred mosquitoes were released into the central courtyard on each of 19

nights of testing.

The mean occupancy rate in rooms treated with Netprotect ITNs vs. untreated
bednets before washing and after one and two washes did not differ (Figure 3.21). There
was no significant difference in mortality at 8 h and 24 h for Netprotect before and after
washing. However, there was a significant difference between Netprotect and untreated

controls for all three treatments (P < 0.0001).

97

 

25%
20% *

 

 

cub
I-r-iD
i—v—ifl

 

 

 

 

 

   

 

 

 

 

 

 

 

 

 

T v I

Untreated Netprotect Untreated Netprotect Untreated Netprotect
Unwashed Unwashed Washed Washed Washed Washed

Mean Percent Occupancy
2% 32' § 5":

 

 

Once Once Twice Twice
Rooms with Untreated vs. Treated Bednets
120 A 120 B

35

      

 

Ileen Percent Mortality ('I.)
8

 

 

[teen Percent lonelity (it)
8

 

 

 

404 40

20‘ b b b 20
om'.'f+'lr...vl"1'.. 0 -. .-
UII- Net. Un- Net- Un. Het- Uri. an. Un- list- mi. in.
WWUWPWWPW ummmmprmmpm
Un- Un- WWWW Un- Un- mmmhed Washed
WWW Once N00 Twice washedwsshstnce Once Twice Twice
Rooms with Untreated 9., Treated Bednets Rooms with Untreated vs. Treated Bednets

Figure 3.21. Occupancy and mortality of Anopheles gambiae females in rooms
treated with Netprotect vs. untreated nets before and after washing. A = 8 h
mortality, B = 24 h mortality. Two hundred mosquitoes were released in the central

courtyard on each of 16 nights of testing.

Mortality after 24 h for Netprotect was remarkably high. No evidence for

avoidance was found for any treatment.

98

DISCUSSION

Performance of the test apparatus

The “Mosquito Hotel” proved effective as a tool for gathering data helpful in
elucidating the mode of action of ITNs at the house-scale. Arranging four equivalent
sleeping rooms around a common courtyard into which a common pool of mosquitoes
was released and having a common set of four sleepers always rotate through each of the
four rooms provided unusually high statistical power for this type of study. Across all
tests where mosquitoes were released only in the courtyard, no significant differences
were ever found in entrance/occupancy rates for sleeping rooms, irrespective of whether
their bednets were treated or untreated with insecticides. In a few cases, however,
mortality rates were variable across sleeping rooms thought to present equivalent
conditions, e.g., Figure 3.8bB. The most likely explanation for this outcome is that we
may have failed to train some of the research assistants to be sufficiently gentle when
they collected and transferred mosquitoes using mouth aspirators. In retrospect, we also
erred in permitting certain assistants to collect mosquitoes from the same room on
successive days, rather than having assistants rotate to successive rooms each day so the
variation each imparted to mosquito mortality was equally distributed across treatments.
These deficiencies were corrected after 24 h mortalities exceeding the 20% acceptable
under WHO (2002) guidelines were obtained in several tests (Figure 3.8bB; Figure
3.103). Further experiments proved the source of this variability in mortality was not

attributable to the choice-test apparatus.

99

E ave openings functioned strongly as inlets, and weakly as outlets

An. gambiae females readily entered sleeping rooms through the 20 cm high x
110 cm wide simulated eave openings at the top of the courtyard walls. Cues emanating
from the sleepers were necessary for high entrance/occupancy rates, as rates of
entry/occupancy for rooms without sleepers were trivial in pilot tests (data not shown).
No significant effect was found when the cave openings were fitted with a restrictor
(Figure 3.6 a and b) whose opening into the sleeping room could be regulated from 20 to
3 cm. Perhaps this is not surprising, given that this apparatus acted as a funnel whose
wide end remained unaltered. We elected to conduct the preponderance of tests using the
20 cm cave opening, because Luo houses always present an cave opening at least this
large unlike in West Afiican studies where entrance is at window level (Asidi et al.,
2004). Ready entrance through eave Openings in the Mosquito Hotel is consistent with
the report of Snow et al. (2007) that An. gambiae encountering a house wall fly upward
against the wall and readily find gaps at its top.

Surprisingly few mosquitoes released into a room housing a sleeper under an
untreated bednet exited when the only portal was the 20 cm eave opening into the
courtyard. This suggests that hungry females are disinclined to fly down a concentration
gradient of host cues even after being repeatedly thwarted from feeding by the physical
barrier of the netting. Host cues appear to be a powerful and long-lasting arrestant. In
retrospect, it would have been informative to have released mosquitoes into a room with
no sleeper to determine whether exit rate was higher without than with host cues. Perhaps
the probability of obtaining a bloodmeal is greater when mosquitoes remain in an

occupied bedroom and wait for the sleeper to exit the bednet than when mosquitoes exit

100

that house and search for another domicile. Likewise, few mosquitoes released into a
room housing a sleeper under a treated bednet exited. This suggests that insecticides
present in that room did not drive mosquitoes out through the cave opening.
I T N mode of action: avoidance vs. lethality

The current data overwhelmingly support lethality as the mode of action of the
ITNs tested and seem to falsify the avoidance hypothesis at the spatial scale of room
entry and exit. No evidence whatsoever was found for reduced entrance rates into rooms
with bednets treated with insecticides vs. rooms with untreated bednets. This outcome
contradicts the conclusions from various hut studies conducted in Western Africa. For
example, Lindsay et a1. (1991) reported that Perrnethrin-impregnated bednets deterred
mosquitoes fiom entering the huts with a degree of deterrency proportional to the dosage
of Perrnethrin. Miller et al., (1991) reported that unwashed Perrnethrin-treated bednet
reduced the number of An. gambiae s.l entering huts by 60%. Difference in methodology
and test conditions may partially explain these discrepancies. West African houses do not
offer the large eave openings for mosquito entry typical of East Afi’ican houses. Rather,
mosquitoes are reported to enter through a series of cracks and slits distributed at various
heights along the walls. Mosquitoes would need to crawl rather than fly through such
small openings. Perhaps, when walking, mosquitoes taste insecticide deposits that
function as repellents or deterrents. On the other hand, mosquitoes flying into a room
with an ITN might not be able to smell pyrethroid insecticides of high molecular weights
and thus very low vapor pressures. Differences in the height of portals used as entrance or
exit traps could strongly influence measures of exophily. Since we wished to understand

the behavior of mosquitoes visiting humans in Luo-type housing, openings for mosquito

101

passage were at wall-top level in all our tests. Additionally, mosquito age and experience
might have some bearing on outcomes of house-entry tests. In our study, all mosquitoes
were naive and foraging for their first bloodmeal. In studies using feral females, the
argument can be made that females surviving encounters with ITNs might avoid
returning to those houses or entering rooms presenting cues now associated with a
negative experience. Finally, some studies of mosquito entry and exit rates using normal
housing conditions may have suffered from problems with accuracy in assessing
avoidance vs. lethality due to difficulties in accurately counting dead mosquitoes. In a
typical Afiican house with dirt floors and the clutter of necessary house contents, it is
very difficult to find dead mosquitoes. Moreover, predators and scavengers like spiders
and ants can be abundant. One of the many advantages of the Mosquito Hotel was that all
floor and wall surfaces were covered by white cloth that minimized the difficulty in
finding mosquitoes, be they dead or alive. Effective anti-predator and anti-scavenger
measures were always in place in our tests. Even so, we typically could not account for
ca. 10 % of the mosquitoes released on any given night, suspecting that they escaped
through tiny cracks and crevices that are difficult to eliminate entirely from any housing
structure. Within the protocol limits specified, we consider that these Mosquito Hotel test
results are the most robust and reliable data produced to date for directly addressing the
question of the relative contribution of avoidance vs. lethality to ITN function for a house
containing human sleepers. Lethality is overwhelmingly supported for all types of ITNs
tested here. This was true irrespective of An. gambiae strain (Kisumu laboratory strain vs.

feral house-caught mosquitoes) or age (4-day and 16-day old female mosquitoes).

102

Slight differences with I T N brand

Where mortality for control rooms fell below or near the 20% WHO (2000)
guidelines, Permanet and Netprotect ITNs yielded higher nightly kill of An. gambiae
females than did Olyset bednets (see Figures 3.8b; 3.16). This outcome is consistent with
the finding of Chapter 2 that mosquitoes alighting on Olyset departed sooner than those
landing on Permanet. Thus, less toxicant might be absorbed per visit on Olyset, given that

the dosages of insecticide loaded into each net type are similar.

Recommendation

Under the housing conditions of Western Kenya, all available data indicate that
ITNs operate as lethal traps. We recommend Netprotect and Permanet ITNs over Olyset,
because of more consistent and higher kill. As mortality from deltamethrin was dosage
dependent (Figures 3.12 and 3.19), we support the high-dosage tactic (Guillet et al., 2001;
Corbel et al. 2004), where an overwhelming dose of deltamethrin is delivered to visiting
mosquitoes so that there are very few to no survivors left in a room presenting a bednet.
Under this approach, ITNs are likely to make huge contributions in reducing human
deaths and suffering due to malaria. However, vector managers will need to be vigilant
and be ready to adjust to counter-adaptations by An. gambiae; experience teaches us that

no pest-control tactic is forever fail-proof.

103

REFERENCES

Aldridge WN. (1990) An assessment of the Toxicological Properties of Pyrethroids and
their Neurotoxicity. Toxicology. 21(2): 89—104.

Binka FN, Kubaje A, Adjuik M, Williams L, Lengeler C, Maude GH, Armah GE,
Kajihara B, Adiamah J H, Smith PG. (1996) Impact of permethrin impregnated
bednets on child mortality in Kassena Nankana District, Ghana: a randomized
control trial. Trop Med Int Health. 1: 147-154.

Bockarie MJ, Service MW, Barnish G, Toure YT. (1995) Vectorial capacity and
entomological inoculation rates of Anopheles gambiae in a high rainfall forested
area of southern Sierra Leone. Trop Med Parasitol. 46: 164-171.

Bogh C, Pedersen EM, Mukoko D, Ouma J. (1998) Permethrin-impregnated bednet
effects on resting and feeding behaviour of lymphatic filariasis vector mosquitoes
in Kenya. Med Vet Entomol. 12: 52-59.

Bondareva NI, Artem’ev MM, Gracheva GV. (1986) Susceptibility and irritability caused
by insecticides to malaria mosquitoes in USSR Part 1 Anopheles pulcherrimus.
Med Parazitol Parazit Bolezni. 6: 52-55.

Brown AWA. (1958) Laboratory studies on the behaviouristic resistance of Anopheles
albimanus in Panama. Bull WHO. 19: 1053-1061.

Caraballo A, Rodriguez-Acosta A. (1999) Chemotherapy of malaria and resistance to
antimalarial drugs in Guayana area, Venezuela. Am J Trop Med Hyg. 61 :120-24.

Casida J E. (1980) Pyrethrum flowers and pyrethroid insecticides. Environmental Health
Perspectives. 34: 189-202.

Chandre F, Darriet F, Duchon S, F inot L, Manguin S, Camevale P, Guillet P. (2000)
Modificiations of pyrethroid effects associated with kdr mutation in Anopheles
gambiae. Med. Vet. Entomol. 14: 81-88.

104

Chandre F, Manguin S, Brengues C, Dossou Yovo J, Darriet F, Diabate A, Camevale P
and Guillet P. (1999) Current distribution of pyrethroid resistance gene (kdr) in
Anopheles gambiae c G V complex from West Africa and further evidence for
reproductive isolation of the Mopti form. Parasitologia. 41 :319-323.

Chang CP, Plapp FW JR. (1983) DDT and pyrethroids: Receptor binding and mode of
action in the housefly pest. Biochem Physiol. 20 (1): 7 6-85.

Chareonviriyaphap, T, Sungvomyothin S, Ratanatharn S, and Prabaripai A. (2001)
Pesticide induced behavioral responses of Anopheles minimus, a malaria vector in
Thailand. J Am Mosq Contr Assoc. 17: 13-22.

Chareonviriyaphap T, Roberts DR, Andre RG, Harlan H, Bangs MJ. (1997) Pesticide
avoidance behavior in Anopheles albimanus Wiedemann, a malaria vector in the
Americas. J Am Mosq Control Assoc. 13:171-183.

Chaudhury DS, Rahman SJ. (1967) Observation on the irritability of susceptible and
resistant strains of A. stephensi (Type) to DDT. Bulletin of the Indian society for
Malaria and Other Communicable Diseases.

Coluzzi, M. (1963) Studies on irritability of DDT to Anopheline mosquitoes WHO/V BC
3321-22.

Curtis C E. (2002) Should the use of DDT be revived for malaria vector control?
Biomedica. 22:455-461.

Cuzin-Ouattara N, Van den Broek AHA, Habluetzel A. (1999) Wide-scale installation of
insecticide-treated curtains confers high levels of protection against malaria

transmission in a hyperendemic area of Burkina F aso. Transactions of the Royal
Society of Tropical Medicine and Hygiene. 93: 473-479.

d’Alessandro U, Olaleye BO, McGuire W, Langerock P, Bennett S, Aikins MK,
Thomson MC, Chain MK, Cham BA, Greenwood BM. (1995) Mortality and
morbidity from malaria in Gambian children after introduction of impregnated
bed net programme. Lancet. 345-479.

Davidson, G (1953) Experiments on the effect of residual insecticides in houses against
Anopheles gambiae and Anopheles funestus. Bull Entomol Res. 44:231-254.

105

Dethier VG, Brown B, Smith CN. (1960) The Designation of Chemicals in Terms of the
Responses they elicit from Insects. J of Econ Entomol. 134-136.

Dong, K, Valles, S M, Scharf, M E, Zeichner B and Bennett, GW. (1998) The
knockdown resistance (kdr) mutation in pyrethroid GV-resistant German
cockroaches. Pestic Biochem Physiol. 60: 195-204.

Fabian MM, Toma T, Tsuzuki A, Susumu Saita S, Miyagi I. (2005) Mark-Release-
Recapture experiments with Anopheles Saperoi (Diptera: Culicidae) in the Yona
Forest, Northern Okinawa, Japan South-East Asian. J Trop Med Public Health.
36:54-63.

Gallup JL, Sachs JD. (2001) The economic burden of malaria. Am J Trop Med Hyg. 64:
85-96.

Gary RE, Foster WA. (2001) Effects of available sugar on the reproductive fitness and
vectorial capacity of the malaria vector Anopheles gambiae (Diptera: Culicidae). J
Med Entomol. 38: 22-28.

Gillies MT, Wilkes TJ. (1981) Field experiments with a wind tunnel on the flight speed
of some West Afiican mosquitoes (Diptera: Culicidae). Bull. Entomol. Res. 71:
65-70.

Gimnig JE, Vulule JM, Lo TQ, Kamau L, Kolczak MS, Phillips-Howard PA, Mathenge
EM, ter Kuile FO, Nahlen BL, Hightower AW, Hawley WAN. (2003a) Impact
of permethrin-treated bed nets on entomological indices in an area of intense year
round malaria transmission. Am J Trop Med Hyg. 68(4): 16-22.

Gimnig JE, Kolczak MS, Hightower AW, Vulule JM, Schoute E, Kamau L, Phillips-
Howard PA, ter Kuile FO, Nahlen BL, Hawley WAN. 2003b. Effect of

permethrin-treated bed nets on the spatial distribution of malaria vectors in
western Kenya. Am J Trop Med Hyg. 68(4): 115-120.

106

Guillet P, N’Guessan R, Darriet F, Traore'-Lamizana M, Chandre F, Camevale P. (2001)
Combined pyrethroid and carbamate “two in one” treated mosquito nets: Field
efficacy against pyrethroid-resistant Anopheles gambiae and Culex
quinquefasciatus. Med Vet Entomol. 15: 105-112.

Gut LJ, Stelinski LL, Thompson DR, Miller JR. (2004) Behavior-modifying chemicals:
Prospects and constrains in IPM, pp 73 — 121, in O. Koul, G. S. Dhaliwal, and G.
W. Cuperus (eds). Integrated Pest Management: Potential Constraints, and
Challenges. CABI Publishing, Cambridge, MA.

Habluetzel A, Diallo DA, Esposito F, Larnizana L, Pagnoni F, Lengeler C, Traore C,
Cousens SN. (1997) Do insecticide treated curtains reduce all-cause child
mortality in Burkina Faso? Trop Med Int Health. 2: 855-862.

Hargreaves K, Koekemoer LL, Brooke BD, Hunt RH, Mthembu J, Coetzee M. (2000)
Anopheles funestus resistant to pyrethroid insecticides in South Africa. Med Vet
Entomol. 14: 181-189.

Hawley WA, Phillips-Howard PA, ter Kuile FO, Terlouw DJ, Vulule J M, Ombok M,
Nahlen BL, Gimnig J E, Kariuki SK, Kolczak MS, Hightower AW. (2003a)
Community-wide effects of permethrin-treated bed nets on child mortality and
malaria morbidity in western Kenya. Am J Trop Med Hyg: 68(4): 121-127.

Hawley WA, ter Kuile FO, Steketee RS, Nahlen BL, Terlouw DJ, Gimnig J E, Shi YP,
Vulule JM, Alaii JA, Hightower AW, Kolczak MS, Kariuki SK, Phillips-Howard
PAN. (2003b) Implications of the western Kenya permethrin-treated bed net

study for policy, program implementation and future research. Am J Trop Med
Hyg 68(4): 168-173.

Hemingway J, Ranson H. (2000) Insecticide resistance in insect vectors of human
disease. Annu Rev Entomol. 45: 371-391.

Hougard JM, Corbel V, N'Guessan R, Darriet F, Chandre Akogbto FM, Baldet T, Guillet
P, Camevale P, Traore-Lamizana M. (2003) Efficacy of mosquito nets treated
with insecticide mixtures or mosaics against insecticide resistant Anopheles

gambiae and Culex quinquefasciatus (Diptera: Culicidae) in Cote d'Ivoire. Bulletin
of Entomological Research. 93: 491- 498.

107

Huang J, Walker ED, Giroux PY, Vulule J, Miller JR. (2005) Ovipositional site selection
by Anopheles gambiae: influences of substrate moisture and texture. Med Vet
Entomol. 19: 442-450

IPCS — International Programme on Chemical Safety (2000b). Perrnethrin. In: Pesticide
Residues in Food. Evaluations 1999. Part II - Toxicological. Joint F AO/VVHO
Meeting on Pesticide Residues (J MPR). Geneva, World Health Organization
(WHO/PCS 00.4):161—200.

Joshi GP, Service MW, Pradhan GD. (1975) A survey of species A and B of the
Anopheles gambiae Giles complex in the Kisumu area of Kenya prior to
insecticidal spraying with OMS-43 (fenitrothion). Annals of Tropical Medicine
and Parasitology. 69: 91-104.

Knols BG, Takken W. (1997) Odour-mediated, host-seeking behaviour of Anopheles
mosquitoes: a new approach. Annals of Tropical Medicine and Parasitology.
91(1): 117-118.

Knols BG, De J ong R, Takken W. (1994) Trapping system for testing olfactory responses
of the malaria mosquito, Anopheles gambiae in a wind tunnel. Med Vet Entomol.
8: 386-388.

Kolaczinski JH, Curtis CF. (2000) Comparison of two alpha-cyano pyrethroids when
impregnated into bednets against a pyrethroid resistant and susceptible strain of
Anopheles stephensi (Diptera: Culicidae) and their F1 progeny. Bulletin of
Entomological Research. 90: 119-123.

Lengeler C. (2000) Insecticide-treated bednets and curtains for preventing malaria
Cochrane Database Syst Rev. 320: CD000363.

Lindsay, SW, Martens WJ M. (1998) Malaria in the Afiican Highlands: Past, Present and
Future. WHO Bull. 76(1): 33-45.

Lindsay SW, Adiamah J H, Miller J E, Armstrong JRM. (1991) Pyrethroid-treated bednet
effects on mosquitoes of the Anopheles gambiae complex in the Gambia. Med
Vet Entomol. 5: 477-483.

108

Lindsay SW, Gibson, M E. (1988) Bednets Revisited Old Idea, New Angle Parasitology
Today, 4: 10.

Lines, J. (1996) Mosquito nets and insecticides for net treatment: a discussion of existing
and potential distribution systems in Afiica. Trop Med Int Health 1: 616-632.

Lines JD, Myamba J, Curtis CF. (1987) Experimental hut trials of permethrin
impregnated mosquito nets and cave curtains against malaria vectors in Tanzania.
Med Vet Entomol. 1: 37-51.

Lockwood, J A, Sparks TC, Story RN. (1984) Evolution of insect resistance to
insecticides: a reevaluation of the roles of physiology and behavior. Bull Entomol
Soc Am. 30:41-51.

Magesa SM, Wilkes TJ, anava AEP, Njunwa KJ, Myamba J, Kivuyo MDP, Hill N,
Lines JD, Curtis CF. (1991) Trial of pyrethroid-impregnated bednets in an area of
Tanzania holo-endemic for malaria 2: Effects on the malaria vector population.
Ada Tropica. 49, 97-108.

Mathenge EM, Gimnig JE, Kolczak M, Ombok M, Irungu L, Hawley WA. (2001) Effect
of permethrin-impregnated nets on exiting behavior, blood feeding success and
time of feeding of malaria mosquitoes (Diptera: Culicidae) in western Kenya. J
Med Entomol. 38: 531-536.

Maxwell CA, Myamba J, Njunwa KJ, Greenwood BM & Curtis CF. (1999) Comparison
of bednets impregnated with different pyrethroids for their impact on mosquitoes
and on re-infection with malaria after clearance of pre-existing infections with

chlorproguanil-dapsone. Transactions of the Royal Society for Tropical Medicine
and Hygiene. 93: 4-11.

Miller J E, Gibson G. (1994), Behavioral response of host-seeking mosquitoes (Diptera:
Culicidae) to insecticide-impregnated bed netting: A New Approach to insecticide
bioassays. J. Med. Entomol. 31(1): 114-122.

Miller, JE, Lindsay SW, Armstrong J RM. (1991) Experimental hut trials of bednets
impregnated with synthetic pyrethroid and organophosphate insecticides for
mosquito control in the Gambia. Med Vet Entomol. 5: 465-476.

109

 

Miller JR, Roelofs WR. (1978) Sustained-flight tunnel for measuring insect responses to
windbome sex pheromone. J. Chem. Ecol. 4: 1 87-98.

Moore A. Tabashnik BE. 1989a. Leg autotomy of adult diamondback moth (Lepidoptera:
Plutellidae) in response to tarsal contact with insecticide residues. J. Econ Entomol. 82:
381-384.

Moore A, Tabashnik BE. (1989b) Leg autotomy: a novel mechanism of protecting against
insecticide poisoning in diamondback moth (Lepidoptera: Plutellidae). J. Econ. Entomol.
82: 1295-1298.

Molyneux DH. (2004) "Neglected" diseases but unrecognised successes-challenges and
opportunities for infectious disease control. Lancet. 364:380-83.

Mouchet J, Cavalie P. (1961) L’irritabilite, vis-a-vis du DDT D ’Anopheles gambiae et
d’afunestus dans le Nord’Cameroun. Riv Malariol. 40 (4-6): 191 -217.

Musawenkosi L, Mabaso H, Sharp B, Lengeler C. (2004) Historical review of malarial
control in southern Afiican with emphasis on the use of indoor residual house-
spraying. Trop Med Int Health. 9(8): 846-856.

Nabarro D, Taylor E. (1998) The "Roll Back Malaria" campaign. Science. 280: 2067—
2068.

Nevill CG, Some ES, Mun’ gala VO, Mulemi W, New L, Marsh K, Lengeler C, Snow
RW. (1996) Insecticide treated bednets reduce mortality and severe morbidity
fi'om malaria among children on the Kenyan coast. Trop Med Int Health. 1: 139-
146.

N’guessan RN, Darriet F, Doannio JMC, Chandre F, Camevale F. (2001) Olyset Net
efficacy against pyrethroid-resistant Anopheles gambiae and Culex

quinquefasciatus after 3 years’ field use in Co’te d’lvoire. Med Vet Entomol. 15:
97-104.

National Pesticide Telecommunications Network (NPTN) Fact Sheets (1998). Pyrethrins
and pyrethroids. http://npic.orst.edu/factsheets/pyrethrins.pdf

110

Oketch BA, Gouagna LC, Killeen GF, Knols BGJ, Kabiru EW, Beier J C, Guiyan Y,
Githure J. (2003) Influence of sugar availability and indoor microclimate on
survival of Anopheles gambiae (Diptera: Culicidae) under semifield conditions in
Western Kenya. J Med Entomol. 40(5): 657-663.

Pauluhn J. (1999) Identification and risk Assessment of Pyrethroids in the indoor
environment. Toxicol Lett. 107: 193 - 199.

Pell JK, Spinney MA, Ward KJ. (1989) Observations on the behavior of adult Anopheles
gambiae encountering residual deposits of lambda-cyhalothrin compared with the
other major classes In: Fourth Annual Confererence of the Society for Vector
Ecology European Region. 1-18.

Phillips-Howard PA, Nahlen BL, Kolczak MS, Hightower AW, ter Kuile FO, Alaii J A,
Gimnig J E, Arudo J, Vulule JM, Odhacha A, Kachur SP, Schoute E, Rosen DH,
Sexton JD, 0100 AJ, Hawley WA. (2003b) Efficacy of permethrin-treated bed
nets in the prevention of mortality in young children in an area of high perennial
malaria transmission in western Kenya. Am J Trop Med Hyg. 68:23-29.

Quinones ML, Suarez ME. (1989) Irritability to DDT of natural populations of the
primary malaria vectors in Colombia. J Am Mosq Control Assoc. 5:56-59.

Rachou RG, Moura LM, Duret JP, Kerr JR. (1963) Experiences with the excito-
repellency test box—model OPS In: Proceedings of the 50th Annual Meeting of
the New Jersey Mosquito Exterminators Association and the 19th Annual
Meeting of the American Mosquito Control Association. 442-447.

Ree HI, Loong KP. (1989) Irritability of Anopheles farauti, Anopheles maculatus, and
Culex quinquefasciatus to permethrin. Jpn J Sanit Zool. 40:47-51.

Roberts DR, Manguin S, Mouchet J. (2002) DDT house spraying and re-emerging
malaria. Lancet. 2000; 356: 330-32.

Roberts, DR, Andre RG. (1994) Insecticide resistance issues in vector-bome disease
control. Am J Trop Med Hyg. 50(6):21-34.

111

Roll Back Malaria Cabinet Project. (2000) The African Summit On Roll Back Malaria,
Abuja, Nigeria Geneva: World Health Organization Nabarro DN, Tayler EM,
1998 The "roll back malaria" campaign. Science. 280: 2067-2068.

Rutledge LC, Echana NM, Gupta RK. (1999) Responses of male and female mosquitoes
to repellents in the World Health Organization insecticide irritability test system. J
Am Mosq Contr Assoc. 1 5: 60-64.

Sachs J, Malaney P. (2002) The economic and social burden of malaria. Nature 415: 680-
685. Scott J A, Brogdon WG, Collins FH (1993) Identification of single specimens
of the Anopheles gambiae complex by the polymerase chain reaction. Am J Trop

Med Hyg. 49: 52-529.

Schellenberg JR, Abdulla S, Nathan R, Mukasa O, Marchant TJ, Kikumbih N, Mushi
AK, Mponda H, Minja H, Mshinda H, Tanner M, Lengeler C. (2001) Effect of

large-scale social marketing of insecticide-treated nets on child survival in rural
Tanzania. Lancet. 357: 1241-7.

Scott J G, Matsumura F. (1981) Characteristics of a DDT-Induced Case of Cross-
Resistance to Perrnethrin in Blattella germanica. Pesticide Biochemistry and
Physiology. 16: 21-27.

Shalaby AM. (1966) Observations on some responses of Anopheles culicifacies to DDT
in experimental huts in Gujarat state, India. Ann Entomol Soc Am. 59:938-944.

Smith TA, Leuenberger R, Lengeler C. (2001) Child mortality and malaria transmission
intensity in Afiica. Trends Parasitol. 17: 145-149.

Snow RW, Omumbo JA, Lowe B, Molyneux CS, Obiero JO, Palmer A, Weber MW,
Pinder M, Nahlen B, Obonyo C, Newbold C, Gupta S, Marsh K. (1997) Relation

between severe malaria morbidity in children and level of Plasmodiumfalciparum
transmission in Africa. Lancet. 349: 1650-1654.

Snow RW. (1987) Do bednets protect against malaria? Lancet. 1: 1493 - 1494.

112

Soderlund DM, Clark JM, Sheets LP, Mullin LS, Piccirillo VJ, Sargent D, Stevens J T,
Weiner ML. (2002) Mechanisms of pyrethroid neurotoxicity: implications for
cumulative risk assessment. Toxicology. 171 :3—59.

Takken W, Knols BG. (1999) Odor-mediated behavior of Afrotropical malaria
mosquitoes. Annu Rev Entomol. 44: 131-157.

ter Kuile FO, Terlouw DJ, Phillips-Howard PA, Hawley WA, Friedman J F, Kolczak MS,
Kariuki SK, Shi YP, Kwena AM, Vulule JM, Nahlen BL. (2003a) Impact of
permethrin-treated bed nets on malaria and all cause morbidity in young children
in an area of intense perennial malaria transmission in western Kenya: Cross-

sectional survey. Am J Trop Med Hyg. 68(4): 100-107.

ter Kuile FO, Terlouw DJ, Kariuki SK, Phillips-Howard PA, Mirel LB, Hawley WA,
Friedman JF, Shi YP, Kolczak MS, Lal AA, Vulule JM, Nahlen BL. (2003b)
Impact of permethrin-treated bed nets on malaria, anemia and growth in infants in

an area of intense perennial malaria transmission in western Kenya. Am J Trop
Med Hyg. 68(4): 68-77.

Threlkeld, SFH. (1985) Behavioral responses in Drosophila melanogaster associated
with permethrin and ectiban, pp 29-36 In: Proceedings of the thirty-second annual
meeting, Canadian Pest Management Society Charlottetown, Prince Island, 24-26
June, 1985 Can Pest Manage Soc.

Tordoir WF, Maroni M, He F. (1994) Health surveillance of pesticide workers. A manual
for occupational health professionals (ICOH—ICPS—WHO, vol. 91).

Ungureanu EM, Theodorescu CN, Ungureanu SE, Emil P, Violeta Z. (1963) Studies on
the resistance of M (d) domestica to insecticides WHO/VECT CONTR/ 17: 8 P
(22 REF) (DPMIAC LOC:WHO SHELF).

USAID. (1997) Proceedings Report Technical Consultation on USAID Infectious
Diseases Strategy December, 16-17, 1997 Washington, DC.

Vijverberg HP, van den Bercken J (1990). Neurotoxicological effects and the mode of
action of pyrethroid insecticides. Crit Rev Toxicol. 21(2):105—126.

113

Vulule JM, Beach RF, Atieli F K, Mount DL, Roberts JM, Mwangi RW. (1996) Long-
terrn use of permethrin-impregnated nets does not increase Anopheles gambiae
permethrin tolerance. Med Vet Entomol.10: 71-79.

Vulule JM, Beach RF, Atieli FK, Roberts J M, Mount DL, Mwangi RW. 1994. Reduced
susceptibility of Anopheles gambiae to permethrin associated with the use of permethrin-
impregnated bed nets and curtains in Kenya. Med Vet Entomol 8:71-75.

Walker ED, Archer WE. (1988) Sequential organization of grooming behaviors of the

mosquito, Aedes triseriatus. Journal of Insect Behavior. 1:97-109.

Webster D, Hill AVS. (2003) Progress with new malaria vaccines Bulletin of the World
Health Organization. 81 (12): 902-909.

WHO. (2000) Manual for indoor residual spraying application of residual sprays for
vector control Geneva, World Health Organization (CDS/WHOPES/GCDPP/3).

WHO. (1995) World Malaria Situation in 1992, Part 1: WHO Weekly. Epidemiol Rec.
42: 309 —- 14 (DPMIAC LOC: WHO).

WHO. (1993) A Global Strategy for Malaria Control. World Health Organization
Geneva.

WHO. (1983), Special programme for Research and Training in Tropical Diseases Sixth
Programme Report Chapter 2: Malaria TDR/PR-6/832-MAL.

WHO. (1970) Insecticide resistance and vector control (17th report of the WHO Expert
Committee on Insecticides) Instructions for determining the irritability of adult
mosquitoes to insecticide WHO Tech Rep Ser 433 : 1 58-163.

WHO. (1963). Tech. Rep. No. 265.

Zaim M, J ambulingam P. (2004) Global insecticide use for vector-bome disease control.
2nd ed. Geneva, World Health Organization
(WHO/CDS/WHOPES/GCDPP/2004.9).

114

   

lllllllllllllllllll