The potential for the emergence of dengue fever in the southwestern United States along the U.S./Mexican border

Material Information

The potential for the emergence of dengue fever in the southwestern United States along the U.S./Mexican border
Hayden, Mary
Publication Date:
Physical Description:
182 leaves : ill. ; 28 cm.


Subjects / Keywords:
Dengue -- Transmission -- Mexican-American Border Region ( lcsh )
Mosquitoes as carriers of disease -- Mexican-American Border Region ( lcsh )
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )


Thesis (Ph. D.)--University of Colorado at Denver, 2003.
Includes bibliographical references (leaves 169-182).
General Note:
Department of Health and Behavioral Sciences
Statement of Responsibility:
by Mary Hayden.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
55609489 ( OCLC )

Full Text
Mary Hayden
B.A., Kiel, Germany, 1977
B.A., Metropolitan State College, 1984
M.A., University of Colorado, 1992
A thesis submitted to the
University of Colorado at Denver
in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
Health and Behavioral Sciences

This thesis for the Doctor of Philosophy
degree by
Mary Hayden
has been approved
David Tracer

Hayden, Mary (Ph D.) Health and Behavioral Sciences
The Potential for the Emergence of Dengue Fever in the Southwestern United
States along the U S./Mexico Border
Thesis directed by Professor Craig Janes
Objective: To examine the impact of climate variability and human-
environmental interaction on the proliferation of Aedes aegypti, the mosquito
vector for dengue fever, into a desert region.
Methods: Three locations will be studied: Tucson, AZ, Nogales, AZ, and
Nogales, Sonora, Mexico over a 3 year study period. Four specific aims are
proposed: Because of local climatic variation, I am evaluating maximum and
minimum temperature and dewpoint temperature at the time and in the places
where mosquito surveillance is undertaken. Second, I am using Ikonos lm
resolution satellite imagery in Ambos Nogales and USGS Digital
Orthoquandrangles in Tucson to identify potential mosquito breeding sites
such as junkyards and over-irrigated sectors and to delineate areas without
digitized street information, and I am constructing GIS-based layers of data
pertaining to utilities access and artificial oviposition sites/surveillance data. I
am conducting field inventories of promising sites in which traps are placed to
document inhibitors and facilitators of mosquito interaction with humans.
Third, in order to establish relevance to seasonal variations in climate, 1 am
conducting mosquito surveillance at 20 sites in each of three study sites at 5
time points relevant to climatic variation. Finally, Aim 4, pertaining to public
health outreach, will use data from Aims 1-3 to collaborate with local public
health officials to formulate appropriate measures to curb mosquito
populations and to educate the local medical community and the public on the
risk of emergent DEN/DHF. Results: Swamp coolers and minimum
temperatures above 69 F are predictive of presence of Aedes aegypti eggs.
This abstract accurately represents the contentofth^bandidate^JhesiSj^
recommend its publication.
Craig J&nes

I dedicate this thesis to my family, and, in particular, John, Liam, Darcy, and
Kerry. Without your continued love and support, this research would not have
been possible. You have been a source of encouragement throughout the course of
study and have patiently listened to both the joys and tribulations at all hours of
the day and night at home and from a distance. Thank you!

First and foremost, I would like to thank my committee, and in particular, Craig
Janes, who believed in me right from the outset. I feel fortunate to have had such
support throughout. David Tracers insights and insistence on rigor have proved
invaluable. Duane Gubler has graciously taken time from a demanding job to
answer innumerable questions and gently nudge me in the right direction. Rafael
Moreno has been dedicated to the success of this project right from the outset, and
Gary Clark has provided critical vector biology field training (with a sense of
humor) to a geographer.
Next, I would like to thank my network of friends including Mary and Frank,
Bev, Lori, Eve, Cec, Taffy and Steve, Mary and Jim, Sneff and Mark, Deb and
Joe, Mickey and Gene, and Cate and Tim to name just a few. Their
encouragement has been unflagging. In addition, my colleagues in Arizona have
been remarkable. Frank Ramberg, Henry Hagedom, and Kathleen Walker have
been untiring in providing me with assistance and friendship. This research would
not have been possible without Cecilia Rosales and her dedication to binational
health. Craig Levy and Andrew Comrie have kindly taken time from their busy
schedules to work with me whenever possible. In addition, Philippe Waterinckx
and Deb Thomas have accompanied me across the border at their own expense
imparting valuable expertise. My cohort classmates have always made time to
listen and offer priceless support and advice because they truly understood.
Additionally, Cristina Luisa has been instrumental in keeping me on track both
with the grant and the dissertation.
Finally, I would like to thank the Office of Global Programs at NOAA, NSF,
EPA, and EPRI for financial support to undertake this study.

1. INTRODUCTION......................................................1
2. BACKGROUND.......................................................11
Pathophysiology of Dengue Fever.................................11
History of Dengue Fever.........................................14
Principal Dengue Vectors........................................19
Temperature and Aedes aegypti Populations.......................29
Precipitation and Aedes aegypti Populations.....................32
Climate Change, Climate Variability, and Vector-borne Disease...34
Geographic Information Systems..................................39
Community Involvement in Vector-Borne Disease Control...........41
3. SOCIAL INEQUALITY AND DISEASE....................................44
United States/Mexico Border Region..............................50
Nogales, Mexico.................................................56
The Emergence of Dengue in the Maquiladora Region...............63
4. METHODS..........................................................67
Study Locales...................................................68
Subjects/Study Sites............................................71
5. RESULTS..........................................................80
Environmental Survey Associations Among Egg Counts and
Independent Variables .........................................108
Container Survey...........................................112
Meteorological Data............................................107
Air Cooling Mechanisms.....................................134

Vegetation Cover...........................................145
Meteorological Data........................................147
Data Collection............................................148
Future Research...............................................154
A. Survey Form...............................................157
B. Container Count...........................................161
REFERENCES CITED......................................................169

1.1 A priori Model of Integrated Factors..................................4
2.1 Range of Aedes aegypti in the Americas...............................17
2.2 World Distribution of Dengue 2000..................................19
4.1 Map of Study Area....................................................69
4.2 Study Sites..........................................................70
4.3 Ambos Nogales with Study Sites in Red................................ 71
4.4 Study Site Locations in Nogales, Mexico..............................78
5.1 June 2002 Egg Count (Tucson, AZ).....................................84
5.2 July 2002 Egg Count (Tucson, AZ).....................................84
5.3 August 2002 Egg Count (Tucson, AZ)...................................85
5.4 September 2002 Egg Count (Tucson, AZ)................................85
5.5 July 2002 Egg Count (Nogales, AZ)....................................86
5.6 August 2002 Egg Count (Nogales, AZ)..................................86
5.7 September 2002 Egg Count (Nogales, AZ)...............................87
5.8 July 2002 Egg Count (Nogales, MX)....................................87
5.9 August 2002 Egg Count (Nogales, 1VDQ.................................88
5.10 September 2002 Egg Count (Nogales, MX)...............................88
5.11 Spatial-Temporal Distribution of Positive Sites by Month.............89
5.12 June 2002 Aedes aegypti Surveillance Results for Ambos Nogales.......91
5.13 July 2002 Aedes aegypti Surveillance Results for Ambos Nogales.......92
5.14 August 2002 Aedes aegypti Surveillance Results for Ambos Nogales.....93
5.15 September 2002 Aedes aegypti Surveillance Results for Ambos Nogales..94
5.16 Spatial Distribution of Positive Sites by Cooling Method............Ill
5.17 Final Results of Study..............................................132
5.18 Site in Tucson......................................................136
5.19 Site in Tucson......................................................137
5.20 Site in Tucson......................................................138
5.21 360 View of Site Above.............................................139
5.22 Site in Nogales, MX.................................................141
5.23 Site in Nogales, AZ.................................................142
5.24 360 View of Site Above.............................................142
5.25 Site in Tucson......................................................143

5.26 Culex Larvae.....................................................144
5.27 Site in Nogales, MX..............................................145
5.28 360 View of Site Above..........................................146

2.1 Temperature and Precipitation Normals for Tucson, AZ 1961 -90..........28
5.1 Description of Dependent and Independent Variables in the Study........81
5.2 Egg Count..............................................................83
5.3 Correlations with 10% and 100% Infusion for July Egg Counts............95
5.4 Correlations with 10% and 100% Infusion for August Egg Counts..........96
5.5 Correlations with 10% and 100% Infusion for September Egg Counts.......96
5.6 Amenities..............................................................97
5.7 Number of Sites Comparing Partial and Full Amenities...................97
5.8 Comparison of Study Sites and Amenities................................98
5.9 Screen Frequency.......................................................98
5 .10 Comparison of Study Sites with No Screen, Partial Screen, and Full Screen 99
5.11 Comparison of Study Sites and Screens..................................99
5.12 Air Cooling Mechanisms................................................100
5.13 Comparison of Study Sites and Air Cooling Mechanisms..................100
5.14 Houses with/without Air Cooling Mechanisms............................101
5.15 Vegetation Frequency..................................................101
5.16 Comparison of Study Sites and Amount of Vegetation....................102
5.17 Percent of Vegetation Coverage........................................102
5.18 Irrigation Frequency..................................................103
5.19 Comparison of Study Sites with/without Irrigation.....................103
5.20 Houses with Irrigation................................................104
5.21 Descriptive Statistics for Cumulative Egg Counts (Amenities).........105
5.22 Descriptive Statistics for Cumulative Egg Counts (Screens)...........106
5.23 Descriptive Statistics for Cumulative Egg Counts (Air Cooling)........106
5.24 Descriptive Statistics for Cumulative Egg Counts (Vegetation)........107
5.25 Descriptive Statistics for Cumulative Egg Counts (Irrigation)........107
5.26 Linear Regression on Total Eggs Found.................................108
5.27 Tests of Between-Subjects Effects.....................................109
5.28 Unconditional Logistic Regression.....................................110
5.29 Tests of Between-Subjects Effects and Containers......................112
5.30 June 2002 Correlations of Maximum Temperatures........................113
5.31 July 2002 Correlations of Maximum Temperatures........................114
5.32 August 2002 Correlations of Maximum Temperatures......................114
5.33 September 2002 Correlations of Maximum Temperatures...................115

5.34 June 2002 Correlations of Minimum Temperatures...........................115
5.35 July 2002 Correlations of Minimum Temperatures...........................116
5.36 August 2002 Correlations of Minimum Temperatures.........................116
5.37 September 2002 Correlations of Minimum Temperatures......................117
5.38 June 2002 Correlations of Average Temperatures...........................117
5.39 July 2002 Correlations of Average Temperatures...........................118
5.40 August 2002 Correlations of Average Temperatures.........................118
5.41 September 2002 Correlations of Average Temperatures......................119
5.42 June 2002 Correlations of Maximum Dewpoint Temperatures..................119
5.43 July 2002 Correlations of Maximum Dewpoint Temperatures..................120
5.44 August 2002 Correlations of Maximum Dewpoint Temperatures................120
5.45 September 2002 Correlations of Maximum Dewpoint Temperatures.............121
5.46 June 2002 Correlations of Minimum Dewpoint Temperatures..................121
5.47 July 2002 Correlations of Minimum Dewpoint Temperatures..................122
5.48 August 2002 Correlations of Minimum Dewpoint Temperatures................122
5.49 September 2002 Correlations of Minimum Dewpoint Temperatures.............123
5.50 June 2002 Correlations of Average Dewpoint Temperatures..................123
5.51 July 2002 Correlations of Average Dewpoint Temperatures..................124
5.52 August 2002 Correlations of Average Dewpoint Temperatures................124
5.53 September 2002 Correlations of Average Dewpoint Temperatures.............125
5.54 Partial Correlations of Temperature and Dewpoint on Log-transformed Egg
Sum (two-tailed).........................................................126
5.55 Partial Correlations of Temperature and Dewpoint on Log-transformed Egg
Sum (one-tailed).........................................................127
5.56 Total Eggs by Average Minimum Temperature for 2002 Surveillance.........128
5.57 Unconditional Logistic Regression Total Eggs Found and Swamp Coolers
and Minimum Temperatures.................................................129
5.58 Summary Regression Model.................................................130
5.59 Model Summary............................................................130
5.60 Analysis of Variance.....................................................131

Dengue fever (DEN) is an arthropod-borne virus that has reemerged in the
Americas and is now pandemic in many urban areas in both the tropics and
subtropics. The principal worldwide vector, Aedes aegypti, has infested the
temperate zone in South America, and its range has recently expanded in the
United States into Arizona and West Texas (Gubler, personal communication
2001.) An estimated one-half of the worlds population live in Aedes aegypti
infested regions, and an estimated fifty to one hundred million people suffer from
DEN annually (Pinheiro and Corber, 1997; Guzman and Kouri, 2003). In 1998,
major epidemics occurred in both Asia and the Americas, with more than 1.2
million reported cases (Gubler, 2002).
Dengue Hemorrhagic Fever (DHF) and Dengue Shock Syndrome (DSS),
serious manifestations of classic DEN, have emerged in the Americas with an
initial epidemic outbreak in Cuba in 1981, and DEN is endemic, at present, in all
but six countries in the Western Hemisphere. Currently DEN causes more illness
and death than any other arbovirus affecting humans (Gubler, 2002, Rosen, 1982);
mortality is low in the presence of optimal medical care, but often high in
developing nations (Gubler, 1997). Case fatality rates can be as high as 40% in
some countries (Gubler, 1988).

The discontinuation of the Pan American Health Organization (PAHO)
eradication program for the Aedes aegypti vector for DEN in the early 1970s has
resulted in documented annual outbreaks of DEN in Mexico beginning in 1978
(Koopman, et al., 1991). A more alarming development in Mexico has been the
emergence of the more serious form, DHF, beginning in 1984 (Briseno-Garcia, et
al., 1996). In 1996, there were 400 laboratory-confirmed cases of DEN in the
Mexican state of Sonora (Engelthaler, et al., 1997) as well as an outbreak along
the U.S./Mexican border in the Mexican state of Tamaulipas (CDC, MMWR,
1996). Outbreaks have occurred as recently as 1999 in both Texas and Mexico in
the border regions with confirmed cases of DEN on both sides of the border and
DHF on the Mexican side (Reiter et al., 2002). While Aedes aegypti is typically a
subtropical/tropical mosquito, its presence has been documented yearly in the
desert regions of Arizona since 1994 (Engelthaler, 1997, Hoeck at al., 2003). Its
geographic range now extends from Nogales, AZ and Nogales, Sonora (the twin
border cities known as Ambos Nogales) northward to Tempe, AZ (personal
communication, Craig Levy, 2001). Although DEN has not yet been documented
in Ambos Nogales, the presence of the vector and outbreaks of DEN in other
regions of Sonora support the notion that DEN will occur in this border region.
Furthermore, the potential exists for DEN to emerge as a health problem in the
United States.

There is a need to examine the impact of land use, climate variability,
water storage practices, and population growth on the proliferation of Aedes
aegypti in a desert climate. Population migration from Mexico to the USA has
been implicated in the recent outbreak of DEN in the Laredo border region of
Texas. The likelihood for similar occurrences in the Nogales/Tucson border
region is high given the general similarities in migration patterns. Three
communities have, therefore, been examined: Tucson, AZ and the twin cities of
Ambos Nogales. This research focuses on the re-invasion of the vector, Aedes
aegypti, into the southwestern United States border region and the multi-factorial
contributors to the potential for outbreaks of dengue and DHF in Mexico/United
States, including the hyperendemicity of dengue serotypes in Mexico, rural-to-
urban migration in northern Mexico and subsequent immigration to the United
States resulting from the implementation of the NAFTA agreement, and the
ensuing social inequity as well as land use change resulting from infrastructural
decline. In addition, the ecology of Aedes aegypti in the Tucson, AZ region has
also been examined with a focus on the greening of Tucson as a factor in the
vectors proliferation. The aforementioned, coupled with possible climatic
variability, serve to increase the potential for the transmission of epidemic DEN
and DHF/DSS and are the focus of the research.
The strength of this research lies in its multidisciplinary approach. For too
long, researchers have approached disease control from either a biological/

entomological or biomedical perspective. Human behavior is a seminal
component in the epidemiology of infectious and parasitic diseases. Behavioral
factors affecting disease transmission are complex and may arise because of a
variety of interconnected issues. Such issues range from ecological changes in the
local environment brought about by land use modifications or climatic variations
to the economic transformation of a region based on global shifts in industrial
models. Although these issues are seemingly disparate, together they may result
in inadvertent human and vector behavioral changes that promote the proliferation
of disease.
Factors in Mosquito Proliferation
in the Sonoran Region
Figure 1.1 A priori Model of Integrated Factors. These are factors believed to be
responsible for the re-invasion of Aedes aegypti into the southwestern United
States/northwestem Mexico.

The model depicted above visually describes the interaction among a
number of factors that are believed to be responsible for the presence of Aedes
aegypti in the Sonoran region. Climate variables impact the use of swamp coolers,
the vegetation of an area, and the use of artificial irrigation which in turn
influence vector survivability. These same climate factors also affect how the land
is used regionally. Because this region is semi-arid, it is necessary to irrigate to
maintain the greening that is evident in Tucson, AZ. Nogales, MX does not have
the necessary resources to allow for a greening of the landscape. In fact, with the
arrival of more and more immigrants from the southern states in Mexico in
response to the availability of work, the browning of the landscape has occurred
as these immigrants settle in squatter settlements or colonias without resources
such as running water, sanitation, or sewage.
This research examines the factors delineated in the previously described
model relating to the proliferation of Aedes aegypti in the Sonoran region. The
addressed research question is to what extent each of the factors presented in the
model impacts the dependent variable, the number of Aedes aegypti eggs, a proxy
for adult mosquitoes. Because of a desire to incorporate all facets of the model
into the research and attempt to determine the relative importance of each factor,
the dissertation has been divided into chapters accordingly.
Chapter 1 is the introduction to the research and provides a model to
delineate the interaction among the myriad variables.

Chapter 2, the background chapter, will delve into the history of DEN, a
disease whose pattern of resurgence in the Americas has followed that of
Southeast Asia in becoming progressively severe. In a few brief decades, with the
initial, post-eradication outbreak of DHF in Cuba in 1981, DEN has re-emerged
throughout much of the Americas in hemorrhagic form with almost yearly
outbreaks throughout the Caribbean, Central, and South America. The
pathophysiology of DEN plays an important role in DHF/DSS, and the theories
surrounding the development of more severe forms of classic dengue are
Chapter 2 also focuses on the principal vectors for DEN in the Americas,
Aedes aegypti and Aedes albopictus. Aedes aegypti, a domesticated mosquito, has
been implicated in most epidemics of dengue in the Americas. However, the
geographic range of Aedes albopictus is expanding, and its potential to become an
important vector of dengue in the Americas is increasing. This section addresses
the literature on Aedes aegypti flight range and dispersal as well as overwintering
capabilities, all factors in transmission potential and risk of dengue. Next, I
review the literature on temperature and precipitation as it pertains to both
extrinsic and intrinsic rates of virus development for the mosquito as well as
seasonality of dengue transmission in particular regions of the world. Because this
study is unique in its examination of local temperature and humidity, it is critical
to understand previous studies which have addressed mosquito and/or dengue

proliferation based on large scale weather data to understand whether future
studies will benefit from the use of local scale information such as that presented
in this research.
A review of the impact of climate variability on vector-borne diseases
with a focus on DEN has been included in this background chapter. There is
considerable debate regarding the role of climate in the ecology of disease,
particularly in the expansion of the geographic range of disease vectors. Climate
change debates began to emerge in the 1980s, and, as a result, the
Intergovernmental Panel on Climate Change (EPCC) was established in 1988.
This multi-disciplinary panel provides national governments and policymakers
with climate information', a group of over 2000 scientists has contributed as lead
authors to the three reports that have been issued so far, in 1991,1996, and 2001.
The last two assessment reports have stand alone chapters on human health and
climate in recognition of the growing importance of this topic (McMichael and
Kovats, 2002) This section will briefly explore the literature surrounding climate
variability and the impact of El Nino Southern Oscillation on vector-borne
In order to effectively manage these data as well as the human behavior
component, the use of GIS has been incorporated into this study. The next section
of Chapter 2 reviews the limitations of previous studies in attempting to define the
habitat of Aedes aegypti and/or the epidemiology of dengue transmission during

an outbreak. For the Ambos Nogales region, I have used Ikonos one meter
resolution satellite imagery to assist in defining an area for which there are no
digital maps available. Although the high resolution technology has not been used
in this study to delineate the scope of an epidemic or the route of epidemic
transmission, it could easily be adapted to do so, allowing for foci of an epidemic
to be determined at a household scale for intervention purposes.
In an era in which a dengue vaccine has yet to be approved, community
efforts to reduce mosquito breeding sites play a major role in the curtailment of
epidemic transmission of dengue. In the final section of Chapter II, I have
reviewed the literature from the past two decades which focuses on the
participatory role communities must play in the self determination of disease
Chapter 3 focuses on the theoretical underpinnings of this dissertation in
recognition of the multi-factorial nature of disease transmission, drawing on
demography, economics, and sociology. Although this research is applied and
quantifiable, some of the factors that have allowed for the proliferation of Aedes
aegypti in this region are the result of infrastructural issues that may be associated
with inequity in the distribution of resources. This research supports a fusion of
ideas underscoring the importance of public health action as an integral part of
society, and, therefore, the subsequent sections address the implementation of the
North American Free Trade Agreement (NAFTA) as a factor in cross-border

migration and the history of both Nogales and the establishment of the
maquiladoras (foreign-owned factories) in response to NAFTA.
Chapter 4, the methods chapter, delineates the data collection procedures.
Because this is a three year study, the results presented in this dissertation are
preliminary. In this first year, I have established sixty sites throughout the study
area of Tucson and Ambos Nogales and collected data from June, July, August,
and September of2002. These data include hourly temperature and dewpoint
temperature (the temperature to which the air must be cooled for saturation to
occur) for each of the 60 sites and mosquito egg counts for a one week period in
each of the study months. In addition, from survey information, I have catalogued
and analyzed the relationship among the following: number of containers capable
of holding water at each site; whether a site has running water, sanitation, and
sewage; whether a site has air conditioning, a swamp cooler, both air conditioning
and a swamp cooler or no artificial cooling mechanism; whether a site is irrigated
or not; whether a site has no screens, some screens, or full screens; and what
percentage of the site is covered by vegetation. Pictures of the sites have been
used to illustrate the results of the analysis. In addition, maps of the research sites
have been provided to demonstrate the geographically stratified nature of the
Chapter 5, the results and discussion chapter, provides the statistical
analysis of the data, samples of the data in graphic form, maps detailing the

spatial and temporal nature of the sites which were positive for Aedes aegypti
eggs, and a discussion of the study as well as conclusions from this preliminary
study. In addition, I have suggested future directions for the research.

Pathophysiology of Dengue Fever
The etiological agents of the DEN and DHF are four antigenically related,
but distinct viruses, DEN 1, DEN 2, DEN 3, and DEN 4 which are classified in
the genus Flavivirus, family Flaviviridae. In addition, there are multiple
genotypes. All four serotypes can cause DEN and DHF, diseases characterized by
sudden onset of fever and headache often accompanied by myalgia, anorexia,
arthralgia, and in the case of DHF, evidence of increased vascular permeability
(Halstead, 1997).
Once a person has been bitten by an infected mosquito, the incubation
period is typically 4-7 days followed by sudden onset of high fever, headache,
retro-orbital pain, myalgia, nausea and vomiting. These acute symptoms last
between 3 and 7 days, but the subsequent convalescence may be prolonged for
weeks (Gubler, 1998). The disease described previously, or classic DEN, is rarely
fatal and generally self-limiting. In more severe cases of DHF/DSS, the initial
phase of the illness is similar to classic DEN. However, at the point the fever
begins to subside, signs of circulatory failure or hemorrhaging may occur. There
is an acute increase in vascular permeability that results in plasma leakage. The

standard treatment for DHF/DSS is fluid replacement therapy which can reduce
fatality levels to less than 1% (Gubler, 1998).
There are two current thoughts on the evolution of classic DEN into
DHF/DSS, viral pathogenesis and immunopathogenesis. The possibility that both
theories are factors in the emergence of DHF/DSS is highly likely. The theory of
viral pathogenesis supports an evolution of the pathogen into highly virulent
strains within a serotype that has increased epidemic potential (Gubler, 1998). It
is possible that highly viremic strains have been selected by incompetent vectors
to ensure propagation in the next generation. Molecular virology has provided
proof of strain variation; however, there is no definitive link between this
variation and severe disease. A recent study investigating the epidemiology of
DEN in Sri Lanka, suggests that the dramatic emergence of DHF lends credence
to the hypothesis suggested in earlier studies by Rosen (1977) and Gubler (1978,
1981) that virus strain is a critical factor. This hypothesis was supported in studies
by Lanciotti et al. (1994) as well, and Messer et al. (2002) noted that small
genotypic changes in the dengue virus may be responsible for the persistent DHF
outbreaks in Sri Lanka.
The second theory arises from twenty-year studies in Thailand which
suggest a relationship between previous exposure and disease severity giving rise
to the antibody dependent enhancement theory (ADE). Primary exposure to one
serotype and subsequent exposure to a second serotype predispose one to severe

disease. The majority of children presenting with DSS in Southeast Asia have had
one or more previous dengue infections (Kliks et al., 1989). Moreover, DHF may
be more frequently observed in children than in adults because of greater
microvascular permeability (Gamble, 2000). Recently, antibody enhancement
theory has been expanded. Hsin Loke et al. (2001) suggest that cross-reactive T
cells play a role in DHF pathogenesis based on a large case-control study in
Vietnam using molecular HLA typing of patients with DHF. Through analysis of
cross-reactivity patterns of CD8 T cell responses to dengue virus, the authors
theorize that tissue damage may ensue in addition to vascular leakage, mediating
DHF pathogenesis (Rothman and Ennis, 1999; Spaulding, et al. 1999).
In addition, children under the age of one often develop severe disease as a
result of passively acquired antibodies from the mother even though the infection
is seemingly primary (Gubler, 1988). The Cuban epidemic of 1977 of classic
dengue resulted from infection with DEN 1, the subsequent outbreak of DHF in
1981 was the result of infection by DEN 2 lending support to the ADE theory or
the theory that sequential infection poses a risk for DHF/DSS (Guzman et al.,
Furthermore, research shows some correlation between age and gender
and disease severity. Before the 1980s, children under the age of fourteen were
more likely to be afflicted by DSS in Asia than adults. Between 1975 and 1978,
reports from Thailand and Indonesia showed few children afflicted with severe

disease to be older than fifteen (Nimmannitya, 1987 and Sumarmo et al., 1983).
However, increasing numbers of adults have been afflicted with DHF in the past
two decades in Malaysia and the Philippines (Gubler, 1978, Guzman, 2002).
There is no correlation between age and disease severity in the Americas
(Pinheiro and Chuit, 1998), although reported outbreaks in Cuba and Venezuela
show the majority of DHF cases were found in children (Kouri et al, 1989 and
Guzman et al., 1999). Females over the age of three were more likely to suffer
from DHF than males (Halstead, Nimmannitya, and Cohen, 1970). The age-
related risk for severe disease may, however, also be a function of an immature
immune system, an area in which more research is necessary (Vaughn, 2000).
Studies have also been published relating ethnicity to disease severity. The case
studies published after the Cuban epidemic of 1981 focus on ethnicity, indicating
that whites are more susceptible to severe disease than blacks. However, closer
examination of this report is needed as the conclusions were based on five
individuals in the study who were seropositive for DHF (Guzman, et al., 1990).
History of Dengue Fever
DEN is believed to have originated in either Asia or Africa and the vector
was initially brought to the Americas on trading ships. The vector was spread on
sailing ships throughout the 17th, 18th, and 19th centuries (Laird, 1989), and the
earliest records of outbreaks in the Americas were the French West Indies in

1635, Panama in 1699 ( Gubler, 1997), and Philadelphia in 1780 documented by
Benjamin Rush (Garrett, 1994). Although clinical descriptions in these epidemics
matched those of dengue, these diagnoses are not definitive given lack of
laboratory confirmed infection.
There is controversy surrounding the origin of the word dengue as well.
The word was first used in the New World in the 1828 epidemic in Cuba, and it is
thought to be an adaptation from Ki-Dinga pepo, Swahili for a disease
characterized by a sudden cramp-like seizure, caused by an evil spirit and
brought to the New World via the slave trade (Gubler, 1997).
Morbidity rates in both U.S. and Japanese soldiers from DEN were high
during World War II, and both countries militaries commissioned dengue studies.
The first successful isolation of the virus occurred in 1943 by Japanese
researchers, but because of publication in an obscure journal, Sabin and
Schlesinger were credited with the first virus isolation in 1944 (Gubler, 1997).
Researchers agree that the proliferation of DEN/DHF/DSS during and
after World War U was brought about by war-related disruption of the existing
ecology of Southeast Asia in that water supplies were destroyed and remnants of
abandoned war materials served as breeding grounds ioxAedes aegypti. In
addition, mass migration because of societal disruptions and movement of armies
allowed populations susceptible to dengue to move in and out of cities and
regions. Conditions were thus ideal for the emergence of DHF/DSS in Southeast

Asia, and in Manila, epidemic DHF/DSS emerged in 1953/54. Sporadic epidemic
outbreaks continued in Thailand and the Philippines throughout the 1950s. By the
1970s, DHF/DSS had become a leading cause of hospitalization and death among
children in several Southeast Asian countries (Gubler, 1997).
In the Americas, although DEN 2 was endemic in the Caribbean Basin,
there were no recorded epidemics from 1946-1963. During this time, because of
the threat of urban yellow fever, the Pan American Health Organization (PAHO)
initiated an Aedes aegypti eradication program and was successful in eliminating
the vector from many Central and South American countries. Unfortunately, the
program was discontinued even though Aedes aegypti had not been completely
eliminated from all countries, and by 1995, the distribution map of the vector
looked strikingly similar to the map from the 1940s, prior to concerted eradication
efforts. Beginning in 1963, there were repeated epidemics of classic dengue in
this region, and in 1977 in Cuba, an epidemic of DEN 1 affected 500,000 people
after a thirty-year period with no apparent DEN viral infection on the island. In
1981, DHF/DSS emerged in epidemic form in Cuba, killing 158 of the 10,312
afflicted with severe disease. Since 1981, there have been yearly (except 1983)
outbreaks of DHF/DSS throughout the Central and South American region
(Pinheiro and Corber, 1997). Brazil suffered one of its largest recorded outbreaks
with 357, 615 cases between January and March 2002. DEN 3 has been the
primary serotype associated with this outbreak with a different and much more

rapid profile of diffusion emerging in many Brazilian states (PAHO,
Reinfestation of Aedes aegypti
1930s v 1970 2002
Figure 2.1 Range of Aedes aegypti in the Americas. Pan American Health
Organization figure depicting the range of Aedes aegypti in the Americas from
1930 to 2002. Source:
Since the mid 1990s in the Caribbean basin, there have been increases in
both the number and severity of dengue epidemics with the co-circulation of all
four serotypes in many countries. Smith and Carter (1983) as well as Reiter and
Sprenger (1987) note that commerce has been one important means of vector
dissemination throughout this region. This, coupled with the lack of

infrastructural support for mosquito control, continued population growth,
urbanization, and lack of access to adequate housing, water, and waste
management, as well as environmental degradation in many areas, has led to
conditions similar to those in Southeast Asia in the 1950s when DHF/DSS began
to emerge in epidemic proportions (Service, 1989). Severe dengue has now spread
to most regions in Southeast Asia and is a major public health problem. A similar
scenario is unfolding in the Americas as is evidenced by the recent epidemic in
Brazil, a country that has seen successive outbreaks since 1986 and where DEN 1,
DEN 2, and most recently in 2001, DEN 3 have been introduced.

World Distribution of Dengue 2000
l~l Areas infested with Aedes aegypti
Areas with Aedes aegypti and dengue epidemic activity
Figure 2.2 World Distribution of Dengue 2000. Source:
Principal Dengue Vectors
In the Americas, the principal vector for DEN, Aedes aegypti of the
subgenus Stegomyia, is a domesticated mosquito that prefers to lay its eggs in
water in artificial containers such as flowerpots, tires, and water storage drums.
Flower holding and water storage containers are productive throughout the year,
however, although indoor vases may produce larvae, because the water is liable to

be changed frequently, they may be less likely to yield pupae or adults. Trash
such as tires, bottles, tin cans, etc. are likely to be productive only during the rainy
season, particularly in a desert climate such as the study locale. It has been noted
(Focks et al., 1981, Focks and Chadee, 1997, Kay et al., 1987) that certain types
of containers are more prolific in producing immature Aedes aegypti than others.
In New Orleans, (Focks, 1981) tires were only 6% of the overall container count,
but were responsible for 25.7% of the immatures. The key containers, containers
in which disproportionate numbers of Aedes aegypti are produced, in Queensland,
Australia were wells and rainwater tanks (Tun-Lin, 1995) with rainwater tanks
supporting 60.4% 63% of the collected immature mosquitoes. The most
productive container in Trinidad (Focks and Chadee, 1997) was the flower pot
with an average standing crop of pupae greater that 30 while the least productive
was the indoor flower vase with an average of fewer than 3 pupae per container.
In Fiji, cans and bottles were found to contain only 2.3% of the immatures.
Although most of these studies focused on the number of immatures in a given
container, Focks and Chadee (1997) concluded that pupal surveys were far more
suitable for determining risk of dengue than the traditional indices House,
Container, and Breteau and the results of these surveys were more pertinent in
directing control operations. In keeping with this focus on pupal surveys,
Montgomery and Ritchie (2002) noted the contribution of roof gutters as key
containers in a study undertaken in Queensland where 1st story gutters were

responsible for 92.3% of the positive gutters. The gutters were productive in both
the wet and dry season producing 50.9% of the pupae in the wet season and
39.3% in the dry season, leading the investigators to underscore the need to
continue to treat gutters in this region.
In an earlier investigation of container productivity in Queensland,
Australia, Tun-Lin (1995) maintained that positive sites, defined as premises in
which one container was positive for Aedes aegypti, were more 3.22 times more
likely to remain positive the following year when compared to negative houses.
This is important in terms of directed control efforts; unfortunately, this study did
not document differing environmental conditions between positive and negative
premises, a potential confounder.
Because Aedes aegypti is well adapted to an indoor environment, and the
female has a short flight range averaging 30 50 meters per day (Rodhain and
Rosen, 1997), it has been implicated in urban epidemics of DEN. Trpis et al.
(1995) showed in a mark-release-recapture study of Aedes aegypti in eastern
Kenya that although marked mosquitoes were recaptured in nearly all houses in
the village, numbers were higher in houses proximate to the release site. A study
on Hainan Island, China, a region where Aedes aegypti is present but where no
dengue transmission has been reported, determined that in areas where houses are
in close proximity to one another, the mean distance traveled (MDT) is small, but
females will fly farther distances if the houses and suitable ovipositing sites are

not adjacent (Tsuda, 2001). The authors concluded that mosquitoes flying from
the center of town would be expected to disperse within a 45 meter range. Other
small-scale mark-release-recapture studies have shown female flight range
between 120 meters in northeastern Mexico (Ordonez-Gonazlaez et al., 2001) and
160 meters in northern Australia (Muir and Kay, 1998). Reiter et al. (1995) have
noted that the female may, however, have a longer flight distance based on
availability of oviposition sites and that mosquitoes may cover distances
measurable in kilometers during the extrinsic incubation period if the same
behavior is exhibited during every gonotrophic cycle. A further study by Edman
et al. (1998) concluded that dispersal is a function of availability of water-filled
containers for oviposition sites, but the authors also noted that the females
seemingly returned to the houses where they had been released even though they
needed to travel to a neighboring site to oviposit. Furthermore, in an earlier mark-
release-recapture study, Wolfinsohn and Galun (1953) showed a flight range of
2.5 kilometers over a 24 hour period in the Negev Desert, Israel.
Related studies of clustering of dengue cases in Puerto Rico indicate that
the disease remained within individual households over the short term (three days
or less) and gradually spread over a wider area in a seven week period (Morrison
et al., 1998). This pattern, also noted by Waterman et al. (1985), may be
indicative of a short flight range of the female with the gradual spread of the

disease resulting from human movement of viremic individuals to other areas and
subsequent infection of female mosquitoes.
These studies have connotations for suburban surveillance of Aedes
aegypti in order to provide a clearer understanding of the local ecology of the
vector, particularly in the desert southwest where limited surveillance has been
undertaken; it may be that the vector will fly greater distances where there is a
dearth of available sites for opposition, particularly during the dry season. This,
then, has implications for the effectiveness of community participation in the
reduction of potential breeding sites as it may serve to shift the focal point of an
epidemic from one neighborhood or town to another, pointing to the need for
concerted, city wide, possibly regionwide, implementation of source reduction
efforts. Based on studies in Puerto Rico, Morrison et al.(1998) noted that focal
spraying may be ineffective because of the rapid temporal and spatial spread of
the disease and that municipal-wide procedures need to be implemented. Focks et
al. (1999) used spatial analysis in Hawaii to determine that insecticide aerosol
applications were ineffective and source reduction efforts needed to be attempted.
Further studies using molecular genetics may provide the tools to indirectly
evaluate dispersal behavior (Walker, 2002) and afford needed data for assessing
Aedes aegypti flight range in a desert environment.
Aedes aegypti is a day biting mosquito preferentially feeding the first few
hours after sunrise and the last few hours before sunset rendering bed-netting

oftentimes ineffectual. Aedes aegypti will, however, feed during the day indoors,
in the shade or when it is overcast. Morlan and Hayes (1958) demonstrated that
peak outdoor feeding activity took place one hour before sunset. Studies from
Trinidad show that although Aedes aegypti is predominantly a diurnal feeder,
particularly in rural areas, in urban areas there was a small nocturnal component
to the landing periodicity at both indoor and outdoor sites. This may be the result
of increased electric light availability in urban areas, and it may indicate an
adaptive modification of behavior (Chadee and Martinez, 2000). Chadee (1988)
suggests that because the vector will bite indoors at night when there is peak
human activity; this has epidemiological significance in that more of the human
body is exposed, facilitating human-vector contact. In addition, Chadee and
Martinez (2000) showed a trimodal pattern of landing with peaks in early morning
and late afternoon, but with a smaller mid-morning component as well in both
urban and rural areas, a factor which may also increase epidemic transmission.
Aedes aegypti is a nervous feeder flitting from person to person, a pattern
that was noted by Macdonald (1956) and Scott (1993) among others. This is a
behavior that facilitates epidemic transmission and may account for clusters of
dengue patients in the same or adjacent households whose illness began at or
around the same time (Halstead et al. 1969). Aedes aegypti is, however, an
inefficient vector of dengue, and there must be high levels of viremia for
transmission and infection from the bitten person. This is important because it is

potentially a selection mechanism ensuring that only highly viremic strains are
transmitted because virus virulence has to be maintained at a high level (Gubler,
Although Aedes albopictus has not yet been definitively implicated in
dengue transmission in the Americas and was previously considered solely a pest
in Hawaii (Focks et al., 1999), it was the vector for an outbreak of DEN in 2001-
2002 in the Hawaiian Islands with a total of 119 confirmed cases as of February 3,
2002 (Hawaii Department of Health, 2002). Earlier studies by Ibanez-Bernal
(1997) determined that Aedes albopictus males collected during an outbreak of
dengue in Reynosa, MX between July and December 1995 had been infected with
dengue serotypes 2 and 3. This was the first record of naturally infected Aedes
albopictus in Mexico or Central America. A recent ProMed posting indicates that
this vector has been detected in larval form for the first time in Nicaragua and is
presumed to be widespread from Mexico to Panama (ProMed, 2003). In other
regions of the world, studies have shown that both Aedes albopictus and Aedes
aegypti have been implicated in the transmission of DEN with Aedes albopictus
involved in a human/mosquito cycle in rural areas and Aedes aegypti in an urban
human/mosquito cycle (Gould, 1970, Chan, 1971, Gubler, 1988).
Rodhain and Rosen (1997) maintain that because of the diversity of its
larval habitats, Aedes albopictus is abundant in both rural and peri-urban locales
such as city parks whereas Aedes aegypti is predominant in the center of large

urban ecosystems. Indeed, the highest incidence of confirmed dengue in the
recent Hawaiian outbreak was found in rural areas of Maui. Although Aedes
albopictus is less domesticated than Aedes aegypti, because of strains that can
enter winter diapause, and the subsequent expansion of its geographic range (-5C
appears to be its current northernmost temperature limit), its role in the
transmission of dengue may increase (Rodhain and Rosen, 1997). Furthermore,
studies have shown that Aedes albopictus has demonstrated the ability to breed
indoors in temperate climate zones (Ishii, 1987). However, because of its
susceptibility to oral infection, Aedes albopictus is likely to transmit disease with
low as well as high levels of viremia, making it a less efficient vector of disease
than Aedes aegypti whose resistance to oral infection suggests that it might favor
furthering viral strains associated with high levels of viremia (Rodhain and Rosen,
1997). Because of Aedes aegyptis preference for human blood meals and its
predilection for ovipositing in containers in close proximity to human dwellings,
it has been implicated in epidemic transmission whereas Aedes albopictus is a less
efficient epidemic vector because the female prefers to oviposit away from human
habitation, thereby decreasing the likelihood of human contact (Gubler, 1988).
Furthermore, egg dessication studies by Juliano et al. (2002) indicate that
Aedes albopictus eggs are less hardy than those of Aedes aegypti, particularly in
regions where humidity is low and temperature is high. In addition, in the same
study in Florida, in areas where both Aedes albopictus and Aedes aegypti have

invaded, occupancy of containers by Aedes aegypti increased with the length of
the dry season and an increase in temperature. Although this may be a means of
ensuring coexistence of competitors in a wet-dry tropical ecosystem, it has
implications for species dominance in a desert ecosystem. Furthermore, Alto and
Juliano (2001) suggest that the prediction of climate change in North America
will limit the range of Aedes albopictus to preclude its establishment in arid areas.
Aedes aegypti diy season survival may be critical to the successful
expansion of populations once the rains commence in areas where there is a
distinct wet and dry season. Because its eggs are relatively dessication resistant,
this may ensure its propagation. Another stratagem for survival noted in a study in
Queensland, Australia (Russell, 2001) may be the utilization of subterranean,
nutrient-rich sites such as septic tanks and manholes that provide insulated, year-
round ovipositing habitats in a region where the mean monthly minimum
temperature is 51.8 F and mean monthly minimum rainfall is less than an inch
during the dry season, a rainfall estimate similar to that in Tucson, AZ during the
winter. In addition, in a study of 24 homes (interior and exterior) in central
Tucson during the dry, cool season between January and March 1999, Botz
(2002) found 17% of the houses had eggs indoors, 21 % outdoors, and 8% had
eggs indoors and outdoors. Of these eggs, 23% of the eggs collected indoors were
viable and hatched within 3 weeks compared to 5% of those collected outdoors.
This is in marked contrast to the study undertaken in Queensland where only 1%

of the eggs in surface locations and 10% in subterranean sites were viable after a
four month period. The difference between indoor and outdoor egg viability in
winter in Tucson may be a function of the higher indoor relative humidity related
to the use of swamp coolers during the dry season and warmer, more stable indoor
temperatures. Studies have yet to be undertaken to determine if the springtime
initial crop of Aedes aegypti is overwintering as adults or hatching from existing
Temperature and Precipitation Normals for Tucson, AZ 1961-90

Temp F Precip inches

High Low Average

Jan 63 38.6 51.3 0.87
Feb 67.8 41 54.4 0.7
Mar 72.8 44.6 58.7 0.72

Table 2.1 Temperature and Precipitation Normals for Tucson, AZ 1961-90

Temperature and Aedes aegypti Populations
Because the influence of temperature on mosquito survival and infective
capacity is an important variable in the study of epidemic disease transmission,
the Nogales/Tucson research project incorporates temperature and dewpoint
temperature data from Ambos Nogales and Tucson. Temperature profoundly
influences the gonotrophic cycle; larval size at pupation affects adult size, and,
consequently, the amount of food available for egg development. A study by
Jetten and Focks (1997) showed that temperature is directly proportional to the
required number of blood meals necessary to complete a gonotrophic cycle; the
warmer the ambient temperature, the smaller the adult mosquito, and, therefore,
the more blood meals required. At 32 C, females will attempt twice as many
blood meals as females at 24 C (Focks et al., 2000). However, in a longitudinal
study by Scott et al. (2000) in Thailand and Puerto Rico, blood feedings increased
with warmer temperatures in Thailand, but not significantly so in Puerto Rico.
The study did corroborate research showing that the body size of Aedes aegypti
decreased with an increase in temperature in both locations and concluded that
temperature, not rainfall, affected population dynamics of Aedes aegypti (Scott et
al., 2000). In addition, there was a relationship between fluctuations in female
abundance and temperature in Thailand but not in Puerto Rico (Scott et al., 2000).
It is also possible that Aedes aegypti may adapt to temperature variation as

discussed in a study undertaken in Queensland, Australia by Tun-Lin, Burkot and
Kay (2000). Smaller adult mosquitoes require more blood meals; therefore,
wanner temperatures increase the opportunities for the spread of infection from
mosquito to humans. Although smaller mosquitoes exhibit less survivability, with
large enough mosquito populations, the threat of DEN transmission exists. The
blood meal stimulates oviposition by the female, and the DEN virus is able to
enter the eggs and be passed on to the next generation of mosquitoes (Monath,
1995). Vertical transmission is likely an evolutionary mechanism allowing the
virus to survive during periods with little or no adult activity such as the winter
months in temperate zones. It is interesting to note that vertical transmission is
more successful with certain serotype strains of dengue (Rodhain and Rosen,
1997) although Tesh (1981) notes that infection rates were quite low in laboratory
experiments. A recent study by Joshi, Mourya, and Sharma (2002), however,
indicates that dengue viruses (DEN 3) persist through successive generations, and
that rates of vertical transmission increased initially in the Fi F2 generations.
The extrinsic incubation period (EIP) or the amount of time necessary for
the DEN pathogen to develop in sufficient quantity to be infective on biting is
also, in part, dependent on temperature; higher temperatures result in a shorter
EEP. Aedes aegypti has been shown to transmit DEN efficiently at temperatures of
20C, but not at temperatures of 16C (Kuno, 1997). In studies done on a
Bangkok strain of Aedes aegypti, the EIP of DEN-2 virus was 7 days when

temperatures were maintained at a range between 32 and 35 C and 12 days at
temperatures <30 C. Also of interest in this study was the variation in EIP
relative to the titer of the mosquito infecting virus doseThe average lifespan of the
female is 8 to 15 days (Rodhain and Rosen, 1997); however, under laboratory
conditions, Aedes aegypti survived between 55 and 100 days with temperatures of
27 1 C, and, once infected with DEN, adult females are infective for their
entire lifespan (Kuno, 1997).
Vector competence is an important component in the ability of Aedes
aegypti to transmit dengue viruses and is defined as the intrinsic permissiveness
of a vector to infection, replication, and transmission of a virus (Bennett et al.
2002). This concept of vector competence dates back to studies undertaken by
Carlos Finlay in 1881 when he hypothesized that yellow fever was transmitted by
mosquitoes, a hypothesis corroborated by Walter Reed and associates in 1900.
Later studies pointed to the role of genetics in the ability of mosquitoes to
transmit arboviruses (Hardy, 1988). Rodhain and Rosen (1997) note that
competence varies among different species as well as among different
geographic populations of the same species. Furthermore, Hardy (1988) states
that in addition to geographic differences, there are intraseasonal and yearly
variations in the proportion of competent and incompetent females capable of
transmitting an arbovirus.

Moreover, there are numerous barriers to vector infection with virus
including midgut infection barrier (MIB), midgut escape barrier (MEB) and
salivary gland barrier; MLBs and MEBs were included in the study undertaken by
Bennett and colleagues. Low MIBs and MEBs are indicative of the increased
potential for transmission. Because these factors may be critical in the
epidemiology of dengue and may be variable for different dengue viruses, in an
effort to focus control efforts, genetic studies are underway to determine if genetic
variation influences the severity of DHF. One such study was undertaken on a
regional scale in Mexico and the United States from among 24 collections of
Aedes aegypti. These studies included field collection from both Tucson, AZ and
Houston, TX as well as numerous sites in Mexico. Aedes aegypti mosquitoes were
infected with dengue 2 virus through oral challenge, and vector competence was
based on the rate of virus dissemination to the head tissues. The vector
competence for Tucson was high at approximately 75% with a midgut infection
barrier of approximately 30% and a midgut escape barrier of approximately 8%
(Bennett et al. 2002).
Precipitation and Aedes aegypti Populations
Jetten and Focks (1997) maintain that rainfall influences the abundance of
Aedes aegypti in certain areas where the breeding sites are discarded containers
such as in southwestern Puerto Rico, but not in other regions such as Bangkok

where the breeding sties are typically manually filled containers designated for
domestic use. Although there is a seasonality to dengue transmission in all regions
of the world, the reasons behind this seasonal transmission are poorly understood.
In a retrospective analysis of 33 years worth of monthly data, there was no
seeming relationship between epidemics of DHF in Bangkok and climate
variability (Reiter, 2001). Earlier studies, however, by Gould et al. (1970) on the
island of Koh Samui in the Gulf of Thailand and Lo and Narimah (1984) in
Malaysia, among others, have noted a rise in dengue incidence coinciding with
the onset of rains and a subsequent increase in larval densities. These studies do
not, however, examine environmental factors that may influence greater
availability of oviposition sites in close proximity to dwellings, etc. On the other
hand, Patz et al. (1998) state that there is no existing body of literature providing
evidence of a correlation between adult survival and rainfall. This conclusion was
also arrived at earlier by Sheppard et al. (1969) in a study of adult Aedes aegypti
populations in Wat, Bangkok whereby adult population fluctuations could not be
correlated with changes in either temperature or rainfall. Others such as Hales
(2002) use humidity or vapor pressure as an indicator of risk; this relationship of
higher humidity to increased transmission is not necessarily valid as exemplified
by instances where transmission is heightened during periods of drought when
water storage becomes a critical factor, and A edes aegypti densities are higher
during the diy season than the wet. A positive correlation was observed between

dengue and the dry season by Eamchan (1989) and Van Peenan (1972). Reiter
(2001), in arguing against a simplistic climate related explanation of potential
increase in incidence, points out that simply because transmission is seasonal does
not mean that warmer temperatures or increased rainfall will enhance
transmission; again, this is because of the complexity of the factors involved in
transmission, including herd immunity, access to amenities such as screens and
air conditioning, viral virulence, and human and mosquito movement among other
Climate Change, Climate Variability, and
Vector-bome Disease
Climate change and climate variability are distinct terms based on time
scale differences. Climate variability refers to mean climate fluctuations that
range from less than yearly to 30 years in duration whereas climate change
generally refers to longer term mean climate tendencies. However, it is important
to note that climate change may impact climate variability, for example, in the
occurrence of stronger El Nino Southern Oscillation (ENSO), a warm phase of
natural oscillation in the sea surface temperature of the southern tropical Pacific
Ocean and seesaw shift in surface air pressure at Tahiti and Darwin, Australia
(NOAA, 2003), and La Nina (the reverse of ENSO characterized by a cooling of
the southern tropical Pacific) events or more extreme weather events such as

extensive drought or flooding (Patz et al., 2000). Although climate is just one
factor among many in the resurgence of disease outbreaks, its influence on the
incidence and distribution of arboviruses is one that cannot be ignored. While a
multitude of societal and demographic factors have been responsible for the re-
emergence of certain vector-borne diseases such as DEN, weather impacts the
seasonality, duration and intensity of epidemics where endemicity is present.
Climatic factors strongly influence not only the development and survival
of the mosquito because of ecological conditions, but also the transmission
dynamics of diseases such as DEN by influencing host and vector behavior.
Existing climate change research and forecasts are based on global scale modeling
with an emphasis on variability in climatic means (McLaughlin et al, 2000); other
models have been proposed attempting to elucidate global scale population
dynamics of Aedes aegytpi such as Hopps (2001) study focusing on long-term
climate patterns driven by precipitation, temperature, relative humidity and solar
radiation parameters and existing mosquito densities. A further model proposed
by Hales et al. (2002) suggests that vapor-pressure is one of the key determinants
of where DEN will extend in the developed world based on climate change.
Monthly averages of vapor pressure, rainfall, and maximum, minimum and mean
temperatures were used along with global circulation models and existing disease
outbreak information to project the risk of DEN in the 2050s and 2080s. Authors
of both studies recognize that their models lack social trend input, and that

addressing issues of demographic and economic changes is critical to the
reduction of disease risk, but Hales et al. are interested solely in providing policy
makers with scenarios to aid in establishing priorities for adaptation to climate
change and improvement of existing strategies for disease reduction. However, in
the aforementioned publication, the data are presented in such a manner as to infer
that climate parameters are the only factors in disease transmission thus belying
the importance of other host/vector relationships such as migration, land use, and
host immune status that are integral to disease transmission.
It is abundantly evident from the literature that social, economic, and
behavioral adaptations to the threat of climate change are imperative to reduce the
risk of disease, and it is implied that areas in the temperate zones most likely to be
impacted by increased or renewed disease resurgence as a result of warming
temperatures are those bordering on endemic areas. This is partially because the
vector is already present and warming temperatures may allow the virus to
develop more rapidly by reducing the extrinsic incubation period, but also
because many of the areas of the developed world with adjacent borders to
developing world countries have similar living standards such as along the
US/Mexico border where back and forth migration and lack of herd immunity in
addition to poverty and inadequate public health service may aid in disease
transmission, particularly on the Mexican side of the border. A recent study by
Reiter et al. (2002) investigates a 1999 outbreak of DEN along the U.S/Mexico

border in Texas where transmission was much higher on the Mexican side than
the U.S. side. Although the number of mosquito-infected containers was higher on
the U.S. side, dengue seropositivity was lower in Laredo, Texas than in Nuevo
Laredo, MX. The authors attribute this to better access to amenities such as air
conditioning and screens which would inhibit human/mosquito interaction and
lifestyles which would tend to keep people indoors in air conditioned rooms.
Shorter term studies have indicated a relationship between ENSO and
DEN in Indonesia, Colombia, French Guiana and Suriname (Gagnon, 2001) and
in many island countries in the South Pacific (Hales et al., 1999). In South
America and Indonesia, national averages for temperature, precipitation, and
hydrologic data were formulated into a time series analysis; these data were used
to show the impact of ENSO on epidemics of dengue. An increased incidence of
DEN in the aforementioned countries in the Americas and of DHF in Indonesia
was always associated with warmer temperatures and, in the Americas, with
drought conditions. In Indonesia, epidemics follow immediately after El Nino
related droughts (Gagnon, 2001). While the necessity of water storage in drought
years may provide one explanation, lack of infrastructure in rapidly developing
urban areas and reduced public health services undoubtedly play a role as well. In
the South Pacific, studies show an increase in dengue epidemics associated with
ENSO, particularly in the larger islands where dengue is already endemic (Hales
et al., 1999). Once again, these papers focus solely on the climate parameters,

ignoring other possible explanations for the resurgence of disease such as strain
variation in addition to demographic and socioeconomic factors.
Models indicate that the areas at greatest risk for change in dengue
transmission potential are the temperate zones where transmission is likely to be
currently limited by minimum temperatures, temperatures which impact both the
virus (EIP) and vector survivability (Martens, Jetten and Focks, 1997). Patz et al.
(1998) also reiterate that the fringe areas are at risk because of the non-immune
status of the population bordering an endemic area. However, these models do not
take into account the myriad factors including demographic and social variables
that allowed for the elimination of dengue from regions in the developed world
where it had once been found (IPCC, 2001).With the advent of air conditioning,
use of window and door screens, and adaptation to indoor activities including
work-related tasks as well as social activities, the risk of DEN declined in the
United States. The border regions continue to present a special challenge in this
respect, nonetheless, as many of the amenities such as air conditioning, which are
taken for granted in other parts of the United States, are lacking. This poverty, as
exemplified by lack of amenities such as window screens, previously not
considered a necessity because of the lack of mosquitoes in the desert, acts to
underscore the vulnerability of the population. Land usage has changed to
accommodate a rapidly increasing population in the border region; the change
ranges from the greening of Tucson, Green Valley, and Tubac to the rapid,

unplanned urbanization and further browning of Nogales, MX. In addition, there
is continual daily movement back and forth across the Nogales border, a factor
which likely aids movement of the vector and will facilitate movement of the
virus in the event of an epidemic. The documentation of these prospective
confounding variables such as land-use change and site specific environmental
change and their incorporation into a model are critical in teasing out the role of
climate variability in the proliferation of the vector. This research provides an
important, little studied, link between Aedes aegypti populations and local climate
in the southern Arizona/northem Sonora area as well as an indication of mosquito
survivability based on human/environmental factors in an area where DEN is not
yet endemic, but the vector is found.
Geographic Information Systems
Medical geography or the geography of disease focuses on the relationship
between disease pathology and the environment in what has been termed spatial
epidemiology (Cromley, 2003; Rushton, 2003; Elliott et al., 2000). Assessment
of the potential for disease emergence has been augmented in recent years by
studies utilizing Geographic Information Systems (GIS) to manage spatial and
temporal data towards predicting high- risk locales. Because resources for
surveillance, prevention, and control of dengue are limited, particularly in the
developing world, it is critical to define the scale at which the aforementioned

surveillance, prevention and control can be conducted. Studies from Puerto Rico,
for example, have indicated the need to apply simultaneous control measures to
an entire municipality because of the rapid spread of dengue both spatially and
temporally during an epidemic in Florida, PR in 1991-92 (Morrison, 1998).
Another study undertaken by Focks et al. (1999) underscores the importance of
spatial analysis in vector control based on studies of the inpact of ultra-low
volume spraying to reduce Aedes albopictus populations in Hawaii and of risk
assessment of dengue focused on the number of Aedes aegypti pupae per person
in Trinidad. From these spatial analyses, Focks et al. were able to recommend
local intervention in Trinidad at the site level to control water-holding containers
in addition to advising against the application of insecticide aerosols on the
military base to control populations of Aedes albopictus in Hawaii.
Other studies such as those by Dale et al. (1998), Freier and Flannery
(1998), Kitron (1998), Moloney et al. (1998), Moncayo, Edman, and Finn (2000),
and Thomson and Connor (2000) have employed the use of GIS and Remote
Sensing in the analysis and integration of spatial elements in the epidemiology of
vector-borne diseases. Most of the remote sensing studies have focused on its use
as a predictive tool for habitat identification that can be used effectively for
predicting many arthropod habitats and some diseases such as malaria (Thomson
et al., 1996). Remote sensing has, for example, been used towards detection of
larval habitats of salt marsh pool mosquitoes such as Aedes vigilax (Skuse) in

Queensland, Australia and of freshwater or transient pool mosquitoes such as
Culex annulirostris (Dale et al., 1998), vectors of Ross River and Barmah Forest
virus. Its use has also been noted for the identification of forest wetland habitats
in Massachusetts where different species of mosquitoes and birds have been
implicated in a cycle of eastern equine encephalomyelitis (Moncayo, Edman and
Finn, 2000). However, because of the peridomestic nature of Aedes aegypti,
remote sensing is of limited value in accurately identifying domestic breeding
sites (Moloney et al., 1998). One of the unique ways in which satellite imagery
has been used in the Tucson/Nogales study is for the identification of individual
houses within the study area in Nogales, MX where digital street and house
information is unavailable and where it would be too labor intensive, time-
consuming, and costly to build a digital database. These individually identified
sites are of interest not only in mosquito surveillance, but they could easily be
expanded to include epidemiological information in the event of an outbreak of
DEN in the region.
Community Involvement in Vector-Borne Diseases Control
Because there is presently no viable vaccine available for dengue
prevention, and ultra-low volume spraying has, on a large scale, been relatively
unsuccessful in mosquito population control, many have advocated both a
bottom-up and top-down approach to reduction of dengue incidence through

community-wide participation in elimination of productive vector breeding sites
(Gubler, 1989) as well as governmental assistance in the reduction of these sites.
Not only is the physical environment critical in determining factors responsible
for the emergence or re-emergence of vector-borne disease, but the role of the
socio-cultural environment must be examined. Behavioral theories of knowledge
and attitudes must be incorporated into disease prevention as well as ongoing
assessment of what is meant by community participation and what types of
participation are successful. Puerto Rico has served as a recent model for
community-based prevention programs (Winch et al., 2002) as has Vietnam (Kay
et al., 2002). Both countries have experienced a major health burden with ongoing
epidemics of DEN, yet have approached prevention with different strategies in
recent years. In Puerto Rico, repeated studies have underscored the difficulty of
implementing sustained behavioral change despite increasing knowledge of
dengue and recognition of the importance of reduction of vector breeding sites in
the prevention of disease (Gubler, 1989); similar encounters have been
documented in the Dominican Republic (Gordon, 1990) and in Mexico (Winch et
al., 1991; Lloyd et al., 1992).
In Puerto Rico, a recent assessment of knowledge and behavior
determined that exposure of children to information about dengue through
classroom discussion, museum exhibits, and posters resulted in higher levels of
correct knowledge about dengue in the treated group than in an untreated group of

children. Although the study may have been skewed in its ability to find a truly
untreated control because of ongoing island-wide information campaigns, the
researchers concluded that children do communicate what they have learned in
school to their parents, but that messages regarding behavioral change need to be
specifically tailored to the parents (Winch et al., 2002). In Vietnam, reduction in
larval indices, viewed as a proxy for the success of behavioral interventions (a
measure that in and of itself may be problematic, however) was achieved through
the use of Mesocyclops (Copepoda), a predacious cyclopoid which successfully
eradicates Aedes larvae from large containers and wells. In addition to the use of
Mesocyclops, community-wide clean-up campaigns using incentives were
initiated such as paying citizens for recyclables that were collected. One of the
more interesting conclusions arrived at in this study was the importance of public
health personnel in relaying a message that was considered to be believable and,
presumably, one upon which villagers acted (Kay et al., 2002).

Political ecology focuses on theories of development and environment,
recognizing that an understanding of the relationship between nature and society
is critical given the growing polarity in world incomes. The interconnectedness
among society, population and the physical and biological environment is crucial
in understanding that a disequilibrium brought about by stress on any one factor
can result in the emergence or re-emergence of disease. All of the aforementioned
factors played a role in the post World War II re-emergence of DEN in Southeast
Asia and its spread to a global scale. The societal disruptions brought about by a
prolonged war resulted in migration to urban areas resulting in human-
environmental interactions that supported the proliferation of the mosquito vector
and the subsequent spread of DEN. Furthermore, acceptance of a multi-causal
role in disease emergence allows for the rejection of the unicausal medical model
of etiology, presupposing that the germ theory holds the key to eradication. Once
the premise of multi-factorial causality is accepted, the importance of the
geographic concept of scale is reinforced with the recognition that factors in the
web of causality range from a submicroscopic scale of disease to globalization.
An analysis of resource use and environmental concerns can be articulated in

local as well as global terms in this research as viewed through climatic
variability at a global and local level and, in conjunction with land use alteration,
its impact on the local transmission of disease. This allows for the linkage among
climate change, land use, and the maquiladora, all of which are intertwined at
both a micro and macro scale, each exacerbating the other in a downward spiral.
Thus, this study emphasizes the basis of disease ecology or the interaction
among humans and the environment in a broader social and economic context
resulting in disease transmission (Mayer, 2000), but with the understanding that
embracing a political or liberation ecology stance requires steps from within the
community to mitigate the potential for dengue emergence.
Political ecology combines the concerns of ecology and a broadly defined
political economy... [which] encompasses the constantly shifting dialectic
between society and land-based resources (Blaikie and Brookfield, 1987:17).
This research focuses on an interconnected ecological and social crisis in the
border regions between the U.S. and Mexico while acknowledging that poverty is
no more a factor in environmental degradation or lack of access to amenities than
is affluence. It is essential to emphasize the myriad of related ills in the chain of
causality and to guard against blaming those who are proximately responsible for
the ensuing environmental degradation. (Peet and Watts, 1996)
In an attempt to move beyond political ecology and refrain from what
Dunn has termed behavior by outsiders (Dunn, 1988), liberation ecology has

evolved as a means of incorporating poststructuralism and discourse theory into
an understanding of geographical place. Since the end purpose of this research is a
sustainable household/community level reduction of mosquito breeding sites and
a decreased likelihood for DEN emergence, the primacy of a grass roots approach
or emphasizing local knowledge is required to achieve the reduction of disease
transmission potential. This can be attempted by the creation of a network of ideas
at a local level that will be sustainable by drawing on indigenous technical
knowledge. However, legitimizing local knowledge, in turn, does not negate the
importance of promoting healthful behaviors or recognizing that indigenous
strategies may be ineffectual or authoritarian (Bebbington, 1996). To this end,
Arturo Escobar (1996) argues for the integration of culture, ecology, and
economy with the recognition that the starting point in this integration is an
articulation of both the consequences of modem development and alternative
ecologically sustainable models.
The correlation between income inequality and health outcomes has been
well substantiated in the literature. Kawachi (1998) has noted the importance of
viewing not only overall countrywide income but also the distribution of wealth
as a determinant of health. In much of the developing world, the disparity
between rich and poor continues to grow. The World Bank estimates that one-
fifth of the worlds population subsists on one dollar a day or less (Brown et al.,
1999), and the subsistence level wages paid to the maquiladora workers has been

well documented. (Pulsipher, 1999; Ramirez, 2000; Marston, 2002) This,
coupled with the dearth of infrastructural amenities, offers a possible explanation
for larger outbreaks of DEN on the Mexican side of the border than on the U.S.
side exemplified by the Laredo/Nuevo Laredo outbreak in 1999.
One means of determining the economic and societal effect of disease is to
assess the impact through disability-adjusted .life years (DALYs). Worldwide
estimates of (DALYs) for DEN and DHF have recently been calculated to
determine the economic impact of these two diseases with an estimated 750,000
DALYs lost each year to DHF. Although the World Bank only looks at DHF,
Meltzer et al. (1998) have established estimates for the impact of DEN. The
burden of DEN and DHF was a loss of 658 DALYs per million population
between 1984-1994 in Puerto Rico. This is a loss of the same magnitude as
tuberculosis, sexually transmitted diseases (excluding HTV), HIV, and malaria
imposed on Latin America and the Caribbean (Gubler and Meltzer, 1999).
Although much of the data must be estimated because of lack of reporting, Gubler
and Meltzer (1999) feel that throughout Central and South America, DEN and
DHF will eventually reach a magnitude of DALYs equal to the burden in
Southeast Asia over the next few decades. Unfortunately, funding from
international agencies has not kept pace with the diseases advance; in 1998 total
funding for DF/DHF was less than US $5 million, and Gubler (2002) notes that

the disparity between funding for malaria (US $84 million in 1998) and DF/DHF
has been even greater in recent years.
In addition to income distribution, social cohesiveness plays a role in well-
being. The more isolated the individual, the higher the mortality rate; this holds
true at a community level as well. The more a neighborhood or community lacks
social cohesiveness, the less likely it is to gamer government support, and
therefore, the more likely it is that its citizenry will succumb to illness and death
whether through infectious disease or increased rates of crime and violence.
Because various components of socioeconomic status (SES) such as income, level
of education, and occupation are influential in determining the course of an
individuals life, they must be examined in a larger social context. In the work
environment, exposure to physical and chemical dangers is key to morbidity and
mortality. In the home environment, social interactions and access to social
supports are determinants of well-being; the fewer community supports available,
and the more stressors an individual perceives, the greater the risk of illness
(Adler et al., 1994). In the colonias of Nogales, although community leaders exist
whose purpose is to ostensibly gamer local government support for the squatters
in the form of provision of city-owned infrastructural amenities such as electricity
and sanitation, many times the squatters pay a fee to build on the land, but the
community leaders do little to improve the circumstances of those from whom
payment is demanded. Other colonia leaders in Nogales have the best interests of

individuals and community in mind providing support, for example, through
provision of oral rehydration packets and domestic violence counseling when
necessary in addition to petitioning the city for services.
Examination of the political economy of poverty requires recognition of
the role it plays in the spread of infectious disease (Farmer, 1996) from an
individual to a community level. At the household level, social and economic
conditions influence behaviors that correlate with health and well-being. Among
the most important factors in the prevention of the transmission of DEN are the
availability of reliable running water, waste disposal, screens and insecticides.
Essential to controlling vector proliferation is ensuring that household heads
understand that the means to this end are entirely within their grasp. The
prevention of larval breeding grounds in proximity to the household entails
covering containers of standing water and disposal of trash. Pontes, et al. (2000)
have shown that source reduction of breeding sites contributes to decreased vector
densities and decreased incidence of DEN. This does not deny the need for
community level intervention; public health services must be made available,
particularly in the arena of mosquito -human interaction through the use of
insecticides and the provision of screens at a household level. Studies have shown
that although broad scale public health measures have been instituted in parts of
southern Mexico, community level compliance with removal of potential breeding
sites for Aedes aegypti is low (Koopman et al., 1991), so the need for household

level intervention remains imperative. Coreil, Whiteford, and Salazar (1997)
have proposed integrating biophysical, social, and culturally constituted
environments to identify disease related behavior in a study of the household role
of DEN transmission in the Dominican Republic. Their study focus on
transmission behaviors such as responsibility for household water procurement
and storage and risk protection behaviors such as house screening highlights the
importance of ascertaining a communitys readiness to participate in health
promoting behaviors as indicative of the achievable level of sustainability within
a given community.
United States/Mexico Border Region
The importance of the geographic concept of place is illustrated in the
contrast between the developed and developing world and is a facet of disease
ecology. Although the United States and Mexico share a 2000 mile common
border, the two countries are worlds apart. The World Bank ranks the United
States in its upper tier economically and Mexico in the second or middle upper
income tier out of four tiers. These rankings are based on integrated information
about life expectancy at birth, gross primary (or secondary) school enrollment,
access to safe water, and gross national product per capita (GDP)

However, the United Nations prefers to use a human development index
(HDI) as a means of measuring a countrys development. This is a composite
index in which life expectancy at birth (based on health and longevity indicators),
education (based on adult literacy and enrollments at all levels in school), and
living standard (based on GDP in terms of purchasing power parity (PPP) are
used as measurements of a countrys achievements. According to this index of
measurement, on a scale of 0 to 1, with 1 the highest possible rank, the United
States ranked 6th out of 162 countries with an HDI of .934 behind 1st ranked
Norway with an index of .936. Mexico ranked 51st with an HDI of .790
This disparity in incomes, life expectancy, and other measures reflected in
the HDI does not necessarily paint a regionally accurate picture. The border
regions of the United States and Mexico defined as 60 miles north and south of
the border, are transitional regions with a unique identity. On the US side, one
finds borderlands characterized by poverty with > 30% of the population at or
below the poverty level (Weinberg, 2003), whereas on the Mexican side some of
the most prosperous regions in Mexico can be found. This binational region has
experienced steady growth in the past two decades in response to the proliferation
of foreign-owned factories or maquiladoras along the Mexican frontier, but also
as a result of other commercial activity with the United States, both of which have
attracted migrants from other states in Mexico and other Central American

countries. The U.S. population of the border region grew by 1.8% annually
between 1993 and 1997, and the Mexican side grew by 4.3% per year during the
same time period. This 2000 mile region is one of the worlds busiest
international borders with an estimated 320 million legal crossings annually and a
population of approximately 11 million people, many of whom continually cross
back and forth to work, shop, visit family and friends and seek medical care
(Weinberg et al., 2003).
The health concerns in this binational region are unique as well although
the residents are not treated as an epidemiologically cohesive unit despite daily
transborder human movement (Doyle and Bryan, 2000). A first step has been
taken to achieve this through the establishment of a Border Infectious Disease
Surveillance (BIDS) project in 1997, a binational endeavor developed by the
Centers for Disease Control and Prevention (CDC), the Mexican Secretariat of
Health, and border health officials to create an active sentinel surveillance system
along the length of the border at 13 clinical sites for Hepatitis (A,B,C,D,E) and
febrile exanthems (measles, rubella, typhus, ehrlicosis, dengue) where case
disease reporting and laboratory disease diagnosis is equivalent on either side
(Weinberg et al. 2003). However, in order to adequately focus on the myriad
public health problems in this region, among which is the potential for the
emergence of DEN, there must be continued binational buy-in and ftmding for
sustained surveillance and prevention of health problems that continue to

disregard the international border as well as expansion of BIDS to include
additional border cities and surveillance for other diseases (personal
communication, Rosales, 2002).
Industrialization and globalization of the economy have long been
recognized as important components of growth in the developing world. Modeled
on the miraculous economic growth that occurred in East Asia in the newly
industrialized countries, in 1965, the Mexican government began to offer
incentives to large corporations that were interested in relocating south of the
U.S./Mexico border under a program known as the Mexican Border
Industrialization Program, maquiladora, or twin plant program. U.S.
corporations were persuaded to move south across the 2000-mile long border so
that they could take advantage of lower wages, fewer taxes, and less stringent
environmental laws. These incentives were based on an economic strategy that
was state-led and encouraged domestic producers to manufacture goods from
imports. However, the Mexican government in the mid-1970s began the
transformation into a neo-liberal export-oriented economy, (Cravey, 1998) and
growth shifted fairly exclusively to a twenty mile strip along the U.S./Mexico

The Mexican debt crisis of 1982 precipitated by plunging oil prices did
little to encourage social equality. The government based growth projections and
investments on expected oil and natural gas income and borrowed heavily,
running up a huge foreign debt. When oil prices fell worldwide, Mexico could not
repay its loans, and an economic crisis ensued. Mexico followed the guidelines set
up by the World Bank, International Monetary Fund (IMF), and the United States
to structurally adjust its economic programs. Mexico instituted currency
devaluation, tax reform and privatization among other measures. Still, these steps
did little to control the ensuing 160% inflation by 1986 and still less to resist the
prescribed privatization of much of the state owned businesses. Mexico became a
model debtor which resulted in a decline in its power and autonomy as it lost the
ability to direct its own domestic economy (Cravey, 1998). Structural Adjustment
Programs (SAPs) such as those instituted by the IMF and World Bank typically
do little to benefit the lower socioeconomic groups; the poor are the first to suffer
and the last to recover. The loss of power and control among those of lower SES
parallels the loss of autonomy by countries forced to submit to IMF and World
Bank reforms.
In January 1994, the North American Free Trade Agreement (NAFTA)
became effective between the U.S., Mexico, and Canada. Since the signing of this
agreement, the number of maquiladoras along the U.S./Mexico border has
skyrocketed from 79 in 1968 to 2,676 by June of 1997 (Pulsipher, 1999). By

October 1999, figures had the number of foreign-owned plants at 4,500 (Preston,
1999). As of October 1999, maquiladora employment had exceeded 1.1 million,
and many cities located along the U.S./Mexico border have doubled their
populations in the last ten years (Shaffer, 1998).
Far from bringing economic prosperity to Mexico, NAFTA has increased
the disparity between the haves and have-nots. Factory jobs have done little to
improve the quality of life for the maquiladora workforce with a $3.50-a-day
minimum wage. While some Mexican workers have gained employment in the
maquiladoras, the countryside and its farmers have suffered as the domestic
market slowly collapses because of a push towards an export-oriented market.
This disintegration of the domestic market forces migration towards the border
areas as the unemployed seek work in the border factories. In 1997, the number of
people seeking jobs increased by 11.5 million while formal employment
accounted for only 4 million new jobs. This coincided with a 70% decrease in real
wages between 1980 and 1997 (Riley, 1999). As migrants move from the
countryside to the urban areas in search of work, squatter settlements have
blossomed. There has been mass migration within Mexico itself; in 1998, 60% of
Tijuanas 1.5 million were migrants from other regions in Mexico. In Baja in
1998, 50% of the population had migrated to the area (Migration News, 1998).
Given such dire economic prospects, it is no surprise that illegal immigration into
the U.S. also continues to increase. In this decade, the U.S. Census Bureau

estimated that in 1990 there were 1.3 million illegal Mexican immigrants in the
U.S., and by 1996, that number had reached 2 million (Riley, 1999). Between
September 1998 and March 1999, the border patrol reported 229,000
apprehensions (Migration News, 1999). Cross-border migration undoubtedly will
continue to rise in response to pull factors on the U.S. side, perhaps increasing the
risk of DEN transmission on the U.S. side through the introduction of the DEN
virus to a non-immune population.
Nogales, Mexico
Nogales, Sonora became a significant port of entiy into the U.S. late in the
nineteenth century when U.S.-built railroads acted to link the two countries. From
WWII until 1964, Nogales operated as a processing center for seasonal migrant
workers entering the U.S. in response to the Bracero Program which was designed
to alleviate labor shortages in the U.S. However, with the cancellation of this
program in 1964, unemployment rose in Nogales. This high rate of
unemployment was a factor in the Mexican governments initial attempt to
develop the border regions by instituting the maquiladora regime. The first step
towards strengthening the federal maquiladora program in Nogales occurred with
the establishment of the Motorola plant in 1967. Motorola was granted a 10-year
tax-free exemption as a model for other companies to invest in the Mexican side
of the border. By 1971, Motorola employed 1,000 Mexicans in a city that now

houses 30 factories and 4,800 factory workers in plants owned by Packard Bell
and Magnavox among others (Cravey, 1998). From 1990 to 1995, the population
of Nogales rose by 25% in response to the continued development of business
(Liverman, 1996) with people migrating not only from rural Sonora, but other
regions in Mexico as well (Cravey, 1998). The population in 2002 is officially at
180,000; however, the true estimated population is 400,000 (personal
communication, Enrique Davis).
Development in Nogales followed the Central Business District (CBD)
model typified in many Central and South American cities. The spatial
framework for the city includes an axis which radiates like a spine from the center
along which many businesses and elite residences are found. The concentric zones
which emanate from the city center are the zone of maturity, home to middle class
residents, the zone of in situ accretion, a transition area from middle to lower
classes, and finally, the peripheral zone, an area of squatter settlements (colonias
populares) which is generally home to the newest migrants to the area.
Oftentimes, these colonias populares are referred to as the disamenity sector or
barrio sector where the lowest socioeconomic groups are forced to live (De Blij,
1999). In Nogales, the newest colonias can be found along the southern
perimeters of the city in response to industrialization in that sector. However, the
landscape is continually in flux as scattered squatter buildings appear overnight in
many of the existing Nogales colonias whether in the vicinity of the maquiladoras

or along the hillsides throughout the city. It is interesting to note that in Nogales,
in contrast to the United States, the newest squatter settlements are forced to build
along the hillsides with little access to any amenities, including lack of available
electricity in some neighborhoods, with, however, the best views, an amenity that
is sought after in the developed world. At the time of my initial set-up of sites in
February 2002, three of the dwellings were without electricity. By the time I
returned in June 2002, all of my sites had electricity although it was not
necessarily supplied directly by the city. Some of the dwelling occupants had
tapped in to their neighbors electrical supply with power cords running across the
unpaved street from one house to the next.
Although Nogales is a typical industrial border city, it most closely
resembles the Asian model in one respect the provision of dormitory housing for
the factory workers. As a result of a massive housing shortage brought about by
large-scale rural-to-urban migration into Nogales, the factory owners built single-
sex dormitories that could accommodate workers for a weekly fee. Although this
provision of shelter appears beneficial, in actuality it allows for the invasion of the
workplace into all facets of the workers lives. There are strict regulations
regarding behavior in the home environment; failure to abide by these rules
invites expulsion from the dormitory and, consequently, from the job. The
corollary to the nuclear family as a result of this type of living environment can
readily be surmised. There is no privacy, and the corporation controls all aspects

of courtship, marriage, and parent-child relationships. Workers are fully aware of
the availability of others to take their places if they dont adhere to corporate
policy and, thus, the corporation controls all events in their lives twenty-four
hours per day, seven days per week. Discipline in the dormitories intrudes on all
social and biological arenas; social costs are high. Pregnancy, for example,
results in automatic expulsion (Cravey, 1998) and the need to seek other housing
arrangements. These factors, once again, play an important role in the continued
expansion of the number of invasiones in the Nogales region.
The workers pay approximately 8% of their weekly wages for space in the
dormitory, in some instances as many as 48 workers share a single room filled
with bunk beds. Because of the close quarters and lack of personal space,
violence, sexual abuse, and theft are commonplace. There is an average
occupancy of fifteen months in the dormitories before workers move to live in
other types of housing. The dormitories are continually filled as owners recruit
workers from all over Mexico through radio advertisement to meet fluctuating
labor demands and replace those who have moved on to either smaller dormitories
or back to over-crowded squatter settlements where there is a semblance of a
nuclear family. In addition, only one of the dormitories in Nogales is close to the
workers place of employment. Other outlying dormitories require a dangerous
and polluted hour-long bus ride. In fact, injury rates from the bus ride to work are
higher than the national average and higher than on-the-job injuries (Cravey,

1998). So, not only are the workers subject to environmental dangers at the job
site, but they must also endure hazardous conditions enroute to and from work.
The workers at the maquiladoras can also be found living in housing that
has been provided by the factory row houses. These domiciles consist of three
to four rooms, typically a living space, bedroom, kitchen and bathroom in which
entire families are living. Clusters of row houses can be found in neighborhoods
such as Calle Mediterrano and San Marcos and are purchased by the inhabitants at
the going rate of between 50,000 and 90,000 US dollars, (personal
communication, RosaElena) and are of substandard construction. However, all the
basic amenities lacking in the squatter settlements such as electricity, sewage,
sanitation, and running water, albeit sporadic, are available. As in the squatter
settlements, because living space is lacking, much of the peripheral area outside
of the house is used for storage resulting in an abundance of containers.
The environmental hazards at the work site have not been well
documented, but Cravey (1998) cites exposure to chemicals, machinery that is
outmoded and unsafe, deficiency of protective clothing and/or equipment, and
stress brought about by excessively long work hours and repetitive motion at an
individual level. A typical work week in the maquiladora environment is 52.5
hours, and a 10 lA hour day is the norm. In addition to individual risk, community
level hazards follow as well in exposure to carcinogens, lack of clean, running
water, and overcrowded and unsanitary living conditions in squatter settlements.

Furthermore, because many of the squatter settlements follow the CBD model,
they are located around the periphery of the city and subject to risk from seasonal
flooding of the Santa Cruz River.
Domestic water use in Nogales, Sonora accounts for 75% of the water
usage, yet only 39% of the population has access to running water 24 hours/day. It
is abundantly evident that because of the lack of running water, large segments of
the population require water storage in containers or drums in and around the
households. Many of these drums are discards from the surrounding factories and
were formerly used to store chemicals. Unfortunately, many of the warning labels
on the drums are in English, and residents who must store their water are exposed
to additional carcinogens (Cravey, 1998). In addition to inadequate supply, the
water that is pumped to residents or delivered by truck to squatter settlements is of
questionable purity. Liverman (1996) cites concern about fecal and coliform
pollution, industrial contamination with heavy metals and sulphur, and
agricultural runoff of nitrates. In the colonias, outdoor toilets are commonly
located near water wells, further raising the threat of contamination. There is
continued concern over the flow of contaminants into the surface water and
underground aquifers throughout the maquiladora region (Warner, 1991).
Furthermore, twenty-five percent of the population is without adequate sanitation
with no means of trash disposal, (Liverman, 1996) providing additional potential
breeding sites for mosquitoes. In 2002, with the ongoing drought in this region,

water was purchased from the city of Nogales, AZ and transferred to the Sonoran
side by trucks at the border. Because the continuing drought conditions in the
region have resulted in a severe water shortage emergency, the city of Nogales,
AZ has again in 2003 donated water to Nogales, Sonora at the cost of
$2.53/1000 gallons ( 2003). This resulted in increased
prices for potable water, and necessitated the eventual transfer of water at a
household level into containers which were not adequately sanitized nor covered
to prevent mosquito breeding.
Environmental degradation is created.. by the rational response of the
poor households to changes in the physical, economic, and social circumstances
in which they define their survival strategies. (de Janvry and Garcia, 1988)
There is a federal health care program available to workers in the northern
industrial cities, but many are forced to use private doctors they can ill afford
because of a dearth of health care providers through the Instituto Mexicano de
Seguro Social (IMSS). Because the IMSS clinics are underfunded and
understaffed, many of the workers cannot afford to take the time off from work to
wait an entire day to see a doctor. This may play a role in ability to detect dengue
in the city of Nogales where clinical access is restricted and the poor are unlikely
to seek medical attention unless they are seriously ill. Furthermore, workers
purportedly have access to federally funded child-care, although in Nogales the
waiting list is years long. Public housing is supposedly also available to workers

through the IMSS, but very few people in Nogales are able to utilize this benefit.
The geographic distribution of these benefits is uneven, and most of the money
that workers and employers pay into the IMSS fund is used to pay for these
services in the old factory regimes and not the new northern border regions
(Cravey, 1998).
Workers in the maquiladora regime in Nogales suffer from inadequate
social services that are inexorably bound to employment. Because there is a
doctor available at the worksite, government sponsored programs have declined,
placing the workers at the mercy of their employers for provision of services.
Thus, in all arenas, the balance of power has shifted to favor the employer, and
with a steady stream of workers moving into the area in search of employment,
the slightest infraction either at the worksite or in the dormitory results in job loss
(Cravey, 1998).
The Emergence of Dengue in the Maquiladora Region
Any examination of causality and infectious disease must be multi-
factorial. Responsible factors include ecological changes, such as those due to
agricultural or economic development or to anomalies in the climate; human
demographic changes and behavior, travel and commerce; technology and
industry, microbial adaptation and change, and breakdown of public health
measures (Farmer, 1999). In the maquiladora environment, all of the

aforementioned factors have led to an increased likelihood of dengue outbreaks
along the U S./Mexico border and in-country transmission in both countries albeit
at a much smaller scale in the United States because of access to amenities such
as air conditioning and screened windows. Economic development has taken
place at such a breakneck pace that infrastructural amenities have not kept abreast
of the changes. In many of the squatter settlements in the border region, running
water is not an available commodity compelling residents to store water close to
their homes. Since Aedes aegypti prefers to breed in artificial containers, water
storage provides a suitable environment for mosquito larvae. Furthermore, lack of
adequate sanitation is a factor in the disposal of used tires and other receptacles in
close proximity to living areas in which Aedes aegypti can breed. In many areas
of Nogales, discarded tires are used to bolster the foundations of the existing
structures in the colonias; although the aforementioned tires are filled with dirt to
weight them and are, thus, unlikely to serve as breeding sites for mosquitoes, tires
are stored near human habitation prior to use and fill with water during the rainy
The globalization of the Mexican market has resulted in the downsizing of
rural farms, and, subsequently, rural-to-urban migration has taken place on a large
scale with an estimated average of 75% of the workers in the maquiladoras
having a migrant history. In border cities like Tijuana, 83% of the workers were
migrants who considered their employment to be transitional resulting in a 12%

monthly turnover as workers moved northward into the Unites States. (Ramirez,
2000) This migration has ensured that migrants from many different dengue
endemic regions in Mexico have moved northward in search of work allowing for
hyperendemicity or the co-circulation of dengue serotypes which in turn raises the
possibility of outbreaks of DHF/DSS along the border. In the last decade, there
have been serologically diagnosed outbreaks of DEN in the border regions on
both the U S. and Mexican sides. Texas has been the site of indigenous
transmission of DEN in the United States with clinically diagnosed disease in
1980, 1986, and 1996 ((CDC, MMWR, 1996). In addition, a recent study
documents significant under-reporting of dengue in the 1999 Laredo, Texas
outbreak due to the lack of recognition of symptoms and disease progression
(CDC, MMWR, 2000). It is interesting to note that Aedes aegypti was
documented in 1994 (Engelthaler, et al. 1997) in Tucson for the first time in
almost 50 years. This is the same year that NAFTA was officially implemented,
opening up cross-border trade and encouraging migration of workers to the border
Given the decline of the social services available to the maquiladora
workers, a hazardous living environment, and lack of basic amenities such as
affordable housing, running water, adequate sanitation, and screens and mosquito
repellent, the border regions remain at risk for the emergence of epidemic DEN.

The international border will provide little protection against outbreaks on the U.S

Research design: The research design is an interdisciplinary, retrospective
and prospective evaluation of land use, climate variability, and amenities access
which have been combined to evaluate the impact of these variables on Aedes
aegypti populations and their survival in the three communities of interest.
General Research Plan: Three inter-related quantitative studies have been
undertaken. 1) The current geographic locations where basic amenities are scarce
to nonexistent/unreliable or where land use and the resultant local environment
have afforded breeding sites for Aedes aegypti, the DEN vector, have been
identified and mapped, and field inventories of inhibitors and facilitators of
human/mosquito interaction have been conducted. 2) Local climate parameters at
the time of surveillance in Tucson, AZ and Ambos Nogales have been analyzed to
determine the impact of potential climate change and present climate on vector
survivability. 3) One season of mosquito surveillance has been conducted during
June 2002 (hot and dry), July, August, and September 2002 (monsoon) in Tucson
and Ambos Nogales in order to empirically test the hypothesized correlations
among amenities access, land use, climate, and mosquito survivability. I intend to
conduct surveillance in January of each year as well since it is climatologically

representative of the cold months of the year. In 2002, because of funding, I
initially set up my sites in February. I will continue to conduct the mosquito
surveillance for the years 2003 and 2004. Through these three related years of
surveillance, collaboration with public health officials on both sides of the
U.S./Mexico border in order to target intervention efforts at communities at
highest risk for DEN, can be facilitated.
Study Locales
The study region is in the southwestern United States/northwestem
Mexico in the Sonoran Desert, an arid region covering 120,000 square miles. This
area is characterized by summer midday temperatures that frequently climb above
100 F and infrequent rainfall with less than 12 inches annually in Tucson and 18
inches annually in Ambos Nogales. There are five seasons in the Arizona Upland
biome: a summer rainy season, autumn, winter, spring, and a foresummer
drought. A bimodal pattern to the rainfall exists with a winter rainy season and the
summer monsoon season. The predominant vegetation is drought resistant cacti
and trees such as creosote, prickly pear, cholla, and mesquite at the lower
elevations-, ocotillo, saguaro, barrel cactus and foothill paloverde on the upper
slopes; and blue paloverde and mesquite in the broader floodplains (Shreve 1951).
Tucson, AZ is located in the southwestern United States at an elevation of
2410 feet with a population of486,699 (U.S. Census, 2000). Nogales, AZ is

located 90 km south of Tucson along 1-19 directly on the U S.-Mexico border at
an elevation of 3865 feet and has a stable population of 25,000 (U S. Census,
2000). Nogales, Sonora is located directly across the border with an official
population of 180,000 and an estimated population of between 350,000 and
Figure 4.1 Map of Study Area
The houses in the study region are varied. In Tucson and Nogales, AZ,
most of the houses are one story constructed of brick, stucco, or adobe. There is a
greater variety to the housing structure of my sites in the Nogales, Sonora area,
particularly in the colonias. Many of the houses in the squatter settlements are
semi-permanent and constructed of a variety of materials including, cardboard,

plywood, and corrugated metal siding. As the families become more established,
the houses are constructed of cinderblock. Where there are no services available,
water is stored outside the dwelling in 55 gallon drums which typically have no
lid. Windows often have curtains in front of them, but no glass or screens. In the
more affluent neighborhoods, cinderblock and stucco are used as construction
Figure 4.2 Study Sites

Figure 4.3 Ambos Nogales with Study Sites in Red. Ikonos 1 meter satellite
imagery showing Ambos Nogales with Study Sites.
Subiects/Study Sites
Participants in the study were recruited in Tucson from among people who
had called in to the University of Arizona in response to an announcement
regarding a mosquito study. This list was utilized to try to geographically stratify
the sites to represent varied sectors of the city to allow for more extensive
coverage than studies previously undertaken by the university. Sites have been set
up in all quadrants of the city as well as central locations. In Nogales, AZ
participants were selected based on information from contacts in the public health

sector as well as the mayors office. In Nogales, MX, the support of the binational
office and Ojicina de Salud Publica was critical in establishing sites. A public
health nurse who is well versed in the politics of all of the neighborhoods enlisted
the support of the leaders in each of the squatter settlements for this study. Either
these community leaders agreed to allow sites to be set up on their land, or they
referred us to willing participants. In addition, homes of some of the public health
doctors and nurses have been used as sites.
In all three study locales, because of the expense of the equipment, sites
had to be selected on the basis of obtaining a secure location for the data loggers
(HOBO ProTemp 8 temperature and relative humidity data loggers by Onset
Corporation). Excluded from the study were sites where potential participants
were unwilling to allow me access to secured sites for the duration of the study.
At the present time, there are 20 sites in Nogales, MX, 18 sites in Nogales, AZ, 1
in Rio Rico, AZ (10 minutes north of Nogales, AZ) 2 in Tubac, AZ (20 minutes
north of Nogales, AZ) and 19 in Tucson, AZ. University of Colorado at Denver
Internal Review Board (IRB) approval was obtained prior to inception of the
study, and consent was verbally obtained from all participants in the study prior to
set-up of equipment.

Because of local climatic variation, maximum and minimum dewpoint and
maximum and minimum temperature parameters have been evaluated at the time
and in the places where mosquito surveillance is undertaken, beginning 7 days
prior to outset of the first jar and continuing for 7 days after the outset of the first
jar. Hourly temperature and dewpoint temperature data have been collected for
analysis for the first year of the study from all 60 sites using data loggers
(HOBOs), providing pertinent data for a more accurate regional-scale climate
assessment. HOBOs have been situated at approximately 4 feet above ground
level in a shaded location on the north side of dwellings. Because of the visibility
of the HOBOs and the possibility of vandalism, the geographic locations have
been pseudo-randomized to coincide with the oviposition traps.
Second, in order to establish a correlation among seasonal variations in
climate, mosquito proliferation and human behavior, one seasons worth of
mosquito surveillance has been conducted, and local environmental conditions
have been inventoried with individual surveys at each of the sites. Mosquito
surveillance has been instituted during the times of maximum rainfall which
coincide with the monsoon season (summer), the driest month because of the
highest potential for water storage in close proximity to dwellings in areas lacking
access to amenities (foresummer drought), and the coldest month (winter) to
determine the potential for Aedes aegypti overwintering as adults in favored

niches. Specifically, the monsoon time period (generally lasting 3 months, July-
Sept) has been targeted with a one-week surveillance during each month (total 3
surveillance periods, 1 each July, August, September), as well as 1 week of
trapping during one of the driest months of the year, May or June, and one week
of trapping during the coldest month of the year, January.
The study uses artificial oviposition sites or ovitraps to extend the
limited surveillance currently underway in the Tucson and Nogales, AZ regions
by local health departments to include areas that have been newly identified
through this work as potential breeding sites using the satellite imagery
techniques and utilities maps as well as local knowledge (people who call in to the
University of Arizonas medical entomology laboratory with mosquito complaints
and public health officials on the Mexican side of the border). Surveillance has
been initiated in the Nogales, Sonora region where geographic stratification has
been attempted in order to assess both sides of the dividing highway (Avenida
Obregon) and areas with and without access to amenities such as dependable
running water, sanitation, and sewage.
Aedes aegypti was trapped for the first time since 1946 in Centers for
Disease Control and Prevention C02 traps in the Tucson region in 1994
(Engelthaler, 1997). Since that time, there have been successful yearly trappings
of Aedes aegypti from the Department of Health Services in Nogales, AZ and
Tucson, AZ. At present, hay infusion oviposition traps are being used, and they

have been adapted for greater retention of the infusion broth, a necessity in the
arid climate of Arizona because of high evaporation rates. Past studies have used
8-12 oz black cups or jars, but the ovitraps used for the current study are quart
jars. The greater ovitrap depth allows for maintenance of enough infusion broth
despite high evaporation to be attractive for ovipositing by Aedes aegypti. The
study ovitraps are quart jars painted glossy black with seed germination paper
inserted in each jar to attract the female mosquito and induce egg-laying. Two
ovitraps are placed side-by-side at each site, one with full-strength broth to attract
the mosquito and the other with a 10% solution. Tropical/subtropical studies have
shown that while mosquitoes are attracted to the 100% solution, they prefer the
10% solution for ovipositing (Fay and Perry, 1965', Reiter, 1991). The hay
infusion broth is brewed by a medical entomologist at the University of Arizona
seven days prior to each study period and used in the ovitraps in all three study
sites. The ovitraps are collected four days after outset, and the germination papers
are stored and returned to the lab at the University of Arizona, where eggs are
counted, hatched, reared, and species are identified.
Initially, the University of Arizona model (Engelthaler, 1997) has been
adopted by dividing Tucson into five sectors: north, east, south, west, and central.
A targeted trapping strategy with pseudorandomization of trap placement within
target areas has been used. The trap sites and positive egg sites for all three areas
have been catalogued using GIS.

Third, I have used Ikonos 1-meter resolution satellite imagery in Ambos
Nogales and USGS quadrangles in Tucson to identify potential mosquito breeding
sites such as junkyards and over-irrigated sectors and to delineate study-site areas
in Nogales, MX where no digitized maps are available. I am constructing GIS-
based layers of data pertaining to utilities access and artificial oviposition
sites/surveillance data. Through spatial and temporal analysis of historic (1990 -
2000) and current remotely sensed land use data, community vulnerability to
DEN will be assessed using a geographic information system (GIS) environment
for signature comparison to detect ecological change. All data are being managed
in a GIS platform allowing for weighting of different variables comprised of
historic and currently collected data to target high risk areas in need of directed
public health intervention.
For both study areas, because I have the support of public health officials
on both sides of the border, I have utilized their on the ground knowledge as
well as that of city officials to verify and improve satellite image interpretation. I
have conducted field inventories and produced photographic documentation of an
area within 50 meters of sites in which traps are placed to assess local factors
which promote or reduce risk of human interaction with mosquitoes. Through the
inventories, I have collected detailed information on the following items:
household waste disposal practices; presence or absence of window and door
screens; existence of potential mosquito breeding containers such as old tires and

discarded bottles; general condition of and construction material used for
dwellings notating gaps between roof and exterior walls, gutters, etc. In addition, I
have noted the presence or absence of air conditioners and swamp coolers, factors
which Reiter et al. (2003) suggested were pertinent to the transmission of dengue
along the Texas/Mexico border in the 1999 outbreak in Laredo/Nuevo Laredo. I
have also performed an ecological assessment of each site to determine whether
0-25%, 26-50%, 51-75%, or 76-100% of the site was covered with vegetation.
Three people assessed each site independently (student from the University of
Colorado, post-doctoral fellow from the University of Arizona, neither of whom
was familiar with the egg count at the sites, and the project director) and
consensus was reached for each site. Whether or not a site was artificially
irrigated using a sprinkler system or handheld hose was also noted.

Figure 4.4 Study Site Location in Nogales, MX. Areas without paved roads are
clearly visable. These typically represent "invasiones" or new squatter
settlements with access to only electricity.
Involvement of public health officials on both sides of the border has been
essential to the success of this study not only to procure study sites, but also to
increase medical community and public health officials awareness of DEN and
work towards the establishment of sustainable programs to reduce Aedes aegypti.
Collaboration with public health officials is an important component of
heightening community awareness of the potential for DEN in the three study
communities. Improving knowledge of DEN serves to enhance reporting of
disease outbreaks and early detection of potential epidemics, complementing the

existing Border Infectious Disease Surveillance Project (BIDS) which has already
begun border surveillance for DEN in regions where the potential for outbreaks
exists. In the Tucson area, with the assistance of University of Arizona medical
entomologists, I have already begun local level education at the study sites and
expanded this to include discussions on the use of Bacillus thuringiensis
israelensis (Bti) as a larvicide. A few of the study participants have begun to use
the larvicide in household, non-removable containers positive for larvae and
pupae, and I have been asked to address neighborhood groups regarding reduction
of breeding sites for both Aedes aegypti and Culex pipiens quinquefasciatus. I
anticipate that through the use of mass media, establishment of neighborhood
watch groups, and elementary school intervention and education, public health
officials messages can be aimed at controlling Aedes aegypti breeding sites and
preventing potential DEN outbreaks.

The study hypothesis being tested is that mosquito egg counts, a proxy for
mosquito population, will be dependent on a number of variables including
climate, number of potential breeding containers at a given site, access to
amenities such as water, sanitation, and sewage, availability of climate cooling
devices in the form of air conditioners or swamp coolers, screened windows and
doors, vegetation or ground cover, and whether or not a site is irrigated.
Access to running water, sanitation, and sewage is a hypothesized factor in
the number of ova because lack of running water or dependable water supplies
would necessitate storing water in close proximity to a dwelling, and, if the water
containers are not tightly closed, they would provide an oviposition site for the
female, just as lack of sanitation often results in an abundance of potential
container breeding sites. Air conditioners may act to reduce mosquito populations
because of decreased humidity and cooler temperatures while swamp coolers may
provide increased humidity in a desert climate, a factor that may enhance
mosquito survivability. Screened windows and doors serve to inhibit mosquito
entrance to the dwelling, a factor in decreased human/mosquito interaction.
Vegetation or ground cover may impart a means of assessing potential mosquito

survivability because of its provision of a protective microclimate against a hot,
dry site in a desert climate regime. In addition, irrigation may enhance vegetation
and/or allow water to pool in existing containers outside of the rainy season, again
providing oviposition sites.
The research question is to what extent does each of these independent
variables affect the dependent variable, the number of Aedes aegypti eggs at a
given site? Table 1 names and describes the variables used in this study.
Variable Description cf Variables
Dependent variable # cf eggs at each of 60 sites Data separated into 10% and 100% solution for each surveillance period
Independent variable climate data Max/Mn/Average temperature and dew point for surveillance period
Independent variable access to amenities Access to elecfricity, water, sanitation, and savage
Independent variable air cooling mechanism Air conditioning, swamp cooler, both or none
Independent variable vegetation Scale of 1-4 for vegetative cover at each site. 1=3.25; 4=>76
Independent variable irrigation Irrigation or none at each site
Independent variable screens No screens, some screens, or fully screened w indaws and doors at each site
Independent variable containers Number of possible breeding containers outdoors at each site
Table 5.1 Description of Dependent and Independent Variables in the Study.
Mosquito surveillance number of eggs per site. In the present study
region, in the absence of dengue case data, egg counts have been used as a proxy

for transmission risk potential. During the summer of2002, only one site out of
sixty was positive for Aedes aegypti eggs in June during an extended period of
drought in which there was no recorded rainfall in the Tucson metro area in either
May or June, continuing a 93 day period with no recorded precipitation.
( This dry period ended on July
9th with the arrival of the monsoon or rainy period. Monsoon onset varies in the
southeastern Arizona region with an average start date of July 3rd and officially
begins when dewpoint temperatures are 54 F or greater for three consecutive
days. The average rainfall during the monsoon is 6.06 inches (1971-2000), and
2002 saw 5.79 inches for the season.
The following chart depicts the number of eggs in both 10% and 100%
solution in all three locales over the 2002 collection season. There is a mean of
85.82 (SD of 128.31; SE 16.6) eggs per household.

City Date 10% 100% Total
Tucson 6/7-6/10 32 31 63
Tucson 7/5-7/8 123 6 129
Tucson 8/7-8/10 671 336 1007
Tucson 9/19-9/22 228 60 288
Nog AZ* 7/5-7/8 163 100 263
Nog AZ 8/7-8/10 412 267 679
Nog AZ 9/19-9/22 304 0 304
Nog MX* 7/5-7/8 100 18 118
Nog MX 8/7-8/10 482 1131 1613
Nog MX 9/19-9/22 667 18 685
Total 3182 1967 5149
Table 5.2 Egg Count. Number of eggs in 10% vs. 100% solution for each locale
in the study area and date of outset of ovitraps. For the month of June, no eggs
were found in Nogales, AZ or Nogales, MX.

Number of eggs CTQ Numberofeggs
June 2002 Ggg Count (Tucson AZ)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
July 2002 Egg Count (Tucson AZ)
Figure 5.2 July 2002 Egg Count (Tucson, AZ)

Number of eggs Number of eggs
Aug 2002 5gg Counts (Tucson AZ)
Sept 2002 ^gg Count (Tucson AZ)
600 -|-
500 -
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Figure 5.4 September 2002 Egg Count (Tucson, AZ)

Number of eggs uQ Number of eggs
July 2002 Egg Count (Nogales AZ)
ire 5.5 July 2002 Egg Count (Nogales, AZ)
Aug 2002 Egg Count (Nogales AZ)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Figure 5.6 August 2002 Egg Count (Nogales, AZ)
O 10%

Sept 2002 Egg Count (Nogales AZ)
z 200
1 2 3 4 5 6 7 0 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
Figure 5.7 September 2002 Egg Count (Nogales, AZ)
July 2002 Egg Count (Nogales MX)
600 -I------------- --- -- - -------------------------
500 -------------------------------------------------------------------
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 10 19 20 21 22 23
Figure 5.8 July 2002 Egg Count (Nogales, MX)

Numberofeggs CTQ Number of eggs
Aug 2002 Egg Count (Nogales MX)
ire 5.9 August 2002 Egg Count (Nogales,
Sept 2002 Egg Count (Nogales MX)
Figure 5.10 September 2002 Egg Count (Nogales, MX)

Spatial-Temporal Distribution of Positive Sites
9 (Tuscea) 17 (Negates)
11 (Teicee) 20 (Negates)

SO Miles
Figure 5.11 Spatial-Temporal Distribution of Positive Sites by Month