ti... .. 2 ..
z 1.: Anna»
C ¢vu L

5
2......

2:5.
F54 \ <
. .

 

... .33,
fiwham.

)

.311. .1 :T:
f: r .u .5 .
.RWHH! fruuw

7 .> .z

.1! .3; ; ..

.. in h--.
. 1.. 2-... .
1.1,... “H.323 El.

21.5.... air. x‘mfrx...

 

 

 

3 «a LIBRARY .
' Michigan State
University

This is to certify that the

thesis entitled

Examining Causal Hechanisns in the
Obesity-Asthma Link: Effect of Body Mass

Index on Imunoglobulin E and Bronchial
Hyperresponsiveness

presented by
Ihuona Uchechi Eneli

has been accepted towards fulfillment
of the requirements for

Masters of Science degree in Epidemiology

W54 (/1

Major professor

Date 31/; [#03

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/CIRC/DateDue.p65-p. 15

EXAMINING CAUSAL MECHANISMS IN THE OBESITY-ASTHMA LINK:
EFFECT OF BODY MASS INDEX ON IMMUNOGLOBULIN E AND
BRONCHIAL HYPERRESPONSIVENESS
By

lhuoma Uchechi Eneli, MD

A THESIS

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

MASTER OF SCIENCE
Department of Epidemiology

2003

ABSTRACT
EXAMINING CAUSAL MECHANISMS IN THE OBESITY-ASTHMA LINK:
EFFECT OF BODY MASS INDEX ON IMMUNOGLOBULIN E AND BRONCHIAL
HYPERRESPONSIVENESS
By

lhuoma Uchechi Eneli, MD

The objective of this study was to explore pathophysiologic mechanisms
underlying the obesity-asthma link, specifically the effect of body mass index on
atopy and bronchial hyperresponsiveness.

The analysis was based on a cohort study of 515 7-9 year old children in Hesse,
Germany. Asthma was assessed based on a history of physician diagnosis,
pulmonary function tests and bronchial challenge test with hypertonic saline
performed at age 7-9 years. Total immunoglobulin E and esinophil count were
determined and body mass index analyzed in quintiles. In multiple linear
regression analysis, high body mass index obtained at four and seven years was
not associated with an asthma diagnosis, bronchial hyperresponsivess or
increased immunoglobulin E. However, high body mass index at age 7 was a risk
factor for exercise-induced bronchospasm (p=0.05). Breastfeeding was
protective for physician-diagnosed asthma (OR= 0.7 95%Cl 0.5-0.9) and
bronchial hyperresponsiveness (OR = 0.4, 95% CI 0.2-0.9). Body mass index
was not associated with increased risk of physician diagnosed asthma or
bronchial hyperresponsivesness .The positive association noted with exercise
induced asthma however suggests a probable chest wall mechanism rather than

an atopy mediated mechanism.

Cepyright by
IHUOMA UCHECHI ENELI
2003

DEDICATION

I would like to dedicate this thesis to my husband, Okey. Without your constant
support and loving encouragement, it would never have started or come to
completion. You are indeed a very special person. To my wonderful children,
Adaeze, Obiora and Emeka, you always reminded me of what was most

important. I love you all.

iv

ACKNOWLEDGMENTS

I would like to recognize the many individuals who have supported and guided
me through the Master's degree program. My sincere appreciation goes out to
Wilfried Karmaus, Dr, Med, MPH, Matt Reeves, DVM, PhD and Richard Honicky,
MD. I would like to thank you for your time and input in helping me complete my
thesis. I am particularly grateful to Dr. Karmaus, for the opportunity to work with
his research data and Betty Elliott for all her clerical assistance.

I would also like to acknowledge the support I received from Marsha Rappley,
MD, my mentor and Department Chair, throughout my course work and
development as a researcher. I thank my colleagues in the General Pediatrics
Division and staff at the MSU Child Health Care Clinic for their understanding
and for tolerating all the schedule changes. Thank you, Kari Chandler for the
”Chai” tea you made for me on especially tough days. They warmed both my
heart and stomach.

Finally, my thanks go out to the staff of the Department of Epidemiology,
especially Nigel Paneth, MD, MPH and Claudia Holzman, DVM, PhD, family

members and friends, for all their support and encouragement.

 

 

 

 

BIB

TABLE OF CONTENTS

 

 

 

 

 

 

 

 

 

 

LIST OF TABLES ‘ vii
LIST OF FIGURES viii
LIST OF ABBREVIATIONS ix
CHAPTER 1

Introduction 1
CHAPTER 2

Literature Review on Obesity and Asthma 12

Conceptual models of causation 35
CHAPTER 3

Data and methods 52
CHAPTER 4

Results 60
CHAPTER 5

Discussion 72
APPENDICIES
Appendix a: Questions on asthma and symptoms 78

Appendix B: Spearman’s correlation coefficients for selected variables- 79

Appendix C: Percentages of asthma symptoms within BMI quintiles
at age 4 years 80

 

BIBLIOGRAPHY 81

 

vi

 

 

 

Tal

Tab

Tabl.

Table

Table 1 .1
Table 1.2

Table 2.1

Table 2.2

Table 2.3

Table 2.4

Table 2.5

Table 4.1

Table 4.2

Table 4.3

Table 4.4

Table 4.5

Table 4.6

Table 4.7

LIST OF TABLES

US prevalence of adult obesity by gender, age and ethnicity -------- 8

 

Classification of obesity (NHLBI/WHO) 10

Epidemiological studies on causal relationship: cross-sectional
studies 1 6

 

Epidemiological studies on causal relationship between asthma
and obesity: cohort studies 20

 

Epidemiological studies on causal relationship between asthma
and obesity: case control studies 23

 

Epidemiological studies on causal relationship: studies examining

 

 

the effect of weight loss 27
Proposed role of third variables in the relationship between asthma

And obesity 33
Characteristics of study population 61

 

Relation of BMI, breastfeeding and family history of atopy to
asthma 63

 

Association between asthma and BMI using 2000 US growth
charts as reference 64

 

Percentages of atopic conditions and asthma symptoms within

 

BMI quintiles for age 7 years 65
Multiple linear regression results for the effect of breastfeeding

and BMI at ages 4 and 7 years on pulmonary function ~-----—------- 68
Effect of BMI on bronchial hyperresponsiveness 69

 

Effect of BMI on immunoglobulin E and eosinophil count --------- 71

vii

Figure 1.1
Figure 1.2

Figure 2.1

Figure 3.1
Figure 3.2

Figure 4.1

LIST OF FIGURES

U.S. asthma attack prevalence per 1,000 population, 1998 ---------- 3

 

Flow-volume loop with a normal FVC maneuver 6

Schematic diagram showing pathophysiology of airway
inflammation and effect of adipocytes and leptin 41

 

Flowchart of the various phases of the study and the data used- 53
Defining study population and subgroups used for analyses ------ 58
Prevalence of asthma, hay fever, exercise-induced bronchospasm

(EIB) and asthma symptoms by body mass index at age 7
years 66

 

viii

BMI
ISAAC
FEV1
FVC

FEV1/FVC

MMF R
FEF50

PER

TLC

BHR

ATS
CDC/NCHS

NHANES
PFT
BRFSS
vao

NHLBI

LIST OF ABBREVIATIONS

Body mass index

lntemational study on allergy and asthma in childhood
Forced expiratory volume in 1 second

Forced vital capacity I

Proportion of forced expiratory volume in one second to
forced vital capacity

Mid maximal expiratory flow rate

Forced expiratory flow at 50% of the FVC maneuver
Peak expiratory flow rate

Total lung capacity

Bronchial hyperresponsiveness

American Thoracic Society

Center for Disease Control and Prevention/National Center
for Health Statistics

National Health and Nutrition Examination Survey
Pulmonary Function Test

Behavioral Risk Factor Surveillance System
World Health Organization

National Heart Lung and Blood Institute

ix

CHAPTER 1

INTRODUCTION

Asthma, a serious, chronic, and potentially life-threatening disease, affects 10-13
million persons in the United States, of whom 5 million are children.1 The
prevalence of asthma has risen by 75% in the last three decades, with an
especially marked increase (160%) in children aged less than five.2 The rising
prevalence transcends age, gender, ethnicity and geographic location but affects
minority ethnic groups and poor inner city populations disproportionately. The
reason for the increase in prevalence is unclear. Increasing environmental
exposures to “synthetic" materials and indoor allergens, decreased childhood
exposures to natural pattern of disease, changing meteorological patterns and
decreased use of aspirin have all been suggested as possible mechanisms. A
parallel increase in obesity has been seen during the same timeframe as asthma
and recently obesity has been hypothesized to be a risk factor for asthma.

The simultaneous rise in the frequency of these two conditions may not be
coincidental?"23 There is evidence that obesity increases the risk of developing
asthma, but our understanding of the obesity-asthma link remains speculative, as
few studies have addressed plausible biological mechanisms. Chest wall
mechanics may explain in part the phenomenon; however, other findings suggest
that factors such as inflammatory mediators, sex hormones, leptin, genes, dietary
fat and gastroesophageal reflux may all play a role.”30 Additionally, both
disorders have been linked independently to breastfeeding, intrauterine growth,

prematurity and birthweightf’"33 It remains unclear if these early childhood

factors exert an influence as antecedent variables. Further complicating the
proposed association are methodologic issues regarding the definition of asthma
and obesity.

Background: Asthma, a major cause of morbidity, diagnosed in 7-15% US
children during childhood, has increased dramatically within the last three
decades. Collectively, people who have asthma experience well over 100 million
days of restricted activity each year, and asthma is believed to be the most
common reason that students miss school. 1' 3‘ Asthma accounts for an
estimated 13 million doctor visits and 200,000 hospitalizations per year in US.
children.34 Figure 1.1 shows the asthma attack prevalence per 1000 population
by age, gender and ethnicity obtained from the National Health interview survey.
Increased rates are shown among children, African Americans and females. In
1990, costs related to asthma were estimated to total $6.2 billion. The projected
direct and indirect cost of asthma in the year 2000 was expected to double to

$14.5 billion.2

Fig 1. Asthma attack prevalence per 1,000 population, 1998

 

Source: CDC] NCHS, National Health lntervlew Survey

Asthma is a heterogeneous disorder marked by gene-environment interaction.
The American Thoracic Society defines asthma as “a disease characterized by
an increased responsiveness of the aiwvays to various stimuli and manifested by
slowing of forced expiration which changes in severity either spontaneously or as
a result of therapy". In the presence of an antigen/antibody complex, mast cell
degranulation releases inflammatory mediators and cytokines like histamine,
tryptase, interleukin-2, interleukin-4, tumor necrosis factor and platelet activating

factors. These mediators then initiate the cascade of bronchoconstriction,

 

hyp

60%

 

rele
the
asll
hisl
fee
fan

De

ep
ex

ar

th

(I)

fit

hypersecretion of mucus, further infiltration of inflammatory cells such as
eosinophils, basophils, neutrophils and macrophages with further cytokine
release, mucosal edema and desquamation of airway epithelium characteristic of
the pathophysiology of asthma. Risk factors important in the development of
asthma include cigarette smoking (active/passive), allergen exposure, family
history/genetic markers, viral and bacterial respiratory illness, early infant
feeding, history of atopy, ethnicity, birthweight less than 25009ms, occupation,
family history of asthma and possibly diet.

Definition of asthma

A validated method for defining asthma is essential for clinical, basic science or
epidemiologic research, yet no universally-agreed upon definition of asthma
exists. A major limitation of extant studies examining the link between asthma
and obesity is the lack of a rigorous case definition for asthma, thus increasing
the likelihood of misclassification bias. Clinical evaluation, questionnaire history
of diagnosis, asthma symptoms, pulmonary function tests and bronchial
hyperresponsiveness are methods that can distinguish between asthmatic and
non-asthmatic patients. Each method is fraught with sensitivity and specificity
problems, especially in children. The lntemational Study on Allergy and Asthma
in Childhood (ISAAC), a multicenter study to examine the prevalence of allergy
and asthma involving 50 countries and 620,000 children, validated questions on
asthma symptoms and diagnosis in 7-8 year and 13-14 year olds using surveys

and video questionnaire. Most questionnaires focus on history and frequency of

wheeze within the preceeding 12 months. The history of wheeze and nocturnal
cough, but particularly wheeze, correlates strongly with a diagnosis of asthma.
Cough is the common primary presenting symptom of asthma but cough alone is
a poor indicator for asthma because most children who cough do not have
asthma. For example, over a 24-hour recording period, the mean frequency of
cough among a cohort of healthy non-asthmatic 8-12 year olds was 11.3 cough
episodes.35

Wheezing also does not provide a definite diagnosis of asthma as it can occur in
smokers, bronchiolitis syndromes and chronic obstructive pulmonary disease.
Clinical evaluation may miss mild cases, be contingent on the diagnostic
threshold of the physician or reflect increased awareness within the population.

L36 reported over-diagnosis of asthma in children based on the

Mackenzie eta
symptom of cough alone, while a recent Swedish study, found the sensitivity of
correctly diagnosing asthma among general physicians was only 59%.37
Conversely, there is considerable overlap in physician diagnosis of similar lung
conditions, e.g., bronchitis, thereby reducing the accuracy of diagnosing asthma.
Pulmonary function measures of obstructive airway disease, e.g., Forced
Expiratory Volume in 1 second (FEV1), Forced Vital Capacity (FVC), Forced
Expiratory Volume/Forced Vital Capacity (FEV1/FVC), Mid Maximal Flow Rate
(MMFR), and Forced Expiratory Flow at 50% of the FVC maneuver, provide an
objective and qualitative assessment of airflow obstruction. FVC is the maximum

amount of air that can be expired with rapid forceful exhalation following a

maximal inspiration. FEV1 is the volume of gas expired over 1 second from the

start of an FVC maneuver, while the MMFR is the average flow in the middle

50% of an FVC maneuver.

 

4.
Ch

 

 

 

O

Flaw (I'M/mud)

«L.
i
ii

 

 

bun-In.

 

 

 

 

 

 

 

0 2 4 6

Figure 1.2 Flow-volume loop with a normal FVC maneuver
(Gregg Ruppel. Manual of Pulmonary Function Tests. 7’" ed. Mosby 1998. page 41.)

The F EV1/FVC ratio is expressed as a percentage and is a strong indicator of
airway obstruction. MMFR and F EF50 are more sensitive for picking up small

airway disease, however these measures have a wide range of normal values.
Figure 1.2 illustrates a flow-volume loop with a FVC maneuver and different

components of the flow loop.

Flow in liters/sec is plotted on the vertical axis while volume in liters is on the
horizontal axis. The two largest FVC and FEV1 values within 200ml or < 5%

difference is used as the reproducibility criteria. In children the 5% criterion is

usually used. Serial measures of pulmonary function rather than a single

measure are more useful. Bronchial challenge testing is used to detect bronchial
hyperresponsiveness and can be very useful in patients with symptoms
sUggestive of asthma but with normal spirometry tests. There is a consensus that
the definition of asthma should focus on the presence of bronchial
hyperresponsiveness (BHR) and paroxysms of ainrvay narrowing, yet BHR is
plagued by poor specificity. Increased reactivity has been reported in non-
asthmatics even in the absence of respiratory symptoms. BHR is also time
consuming and more invasive than other methods; thus it is impractical for most
epidemiologic studies. To date, no studies have examined the obesity-asthma
link using a combination of clinical, physiologic and questionnaire methods to
define asthma.

Body mass index and obesity: The National Health and Nutrition Examination
Survey (NHANES III 1988-1994) estimates that a quarter of children ages 6-17
years and more than 50% of adults in the United States are overweight.”39 The
prevalence of obese children doubled between the 19705 and 19903.39 Obesity
now accounts for about 7% of the US. healthcare budget, with $68 billion spent
on direct medical expenses.40 Obesity related hospitalizations for children
increased between 1979-1981 to 1997-1999, with tripling of the hospital costs
from $35 million to $127 million.41 Table 1.1 provides prevalence rates for obesity
from 1991-2000 from the Behavioral Risk Factor Survelliance Survey (BRFSS), a
telephone survey carried out in all the states. The prevalence has risen across all

categories especially African-Americans and Hispanic populations.

 

Tal

Table 1.1 US prevalence of adult obesity (approximately 301bs overweight)
by gender, age and ethnicity

 

 

 

. Percent Obese
Characteristics BRFSS data by year:
1991 1995 ‘ 1998 1999 2000

Total 12.0 15.3 17.9 18.9 19.8
Gender

Men 11.7 15.6 17.7 19.1 20.2

Women 12.2 15.0 18.1 18.6 19.4
Age groups

18-29 7.1 10.1 12.1 12.1 13.5

30-39 11.3 14.4 16.9 18.6 20.2

40-49 15.8 17.9 21.2 22.4 22.9

50-59 16.1 21.6 23.8 24.2 25.6

60-69 14.7 19.4 21.3 22.3 22.9

>70 11.4 12.1 14.6 16.1 15.5
Race, ethnicity

White, non Hispanic 11.3 14.5 16.6 17.7 18.5

Black, non Hispanic 19.3 22.6 26.9 27.3 29.3

Hispanic 11.6 16.8 20.8 21.5 23.4

Other 7.3 9.6 11.9 12.4 12.0
‘CDC/NCHS

Obesity is a complex, multifactorial disorder resulting from genetic, environmental
and psychogenic factors that impact energy intake and expenditure.
Complications of obesity include hyperlipidemia, type 2 diabetes, and increased
incidence of cardiovascular disease, stroke, cancers and mortality. In children
and adolescents, the sequelae were once felt to be predominantly psychosocial,
stemming from discrimination, low self-esteem and accelerated maturation.
However, with the recent rise in prevalence, medical complications previously
seen only in adults, such as impaired glucose metabolism, non-insulin dependent
diabetes, dyslipidemia, and hypertension, now occur in obese children and

adolescents.

Body mass index (BMI), the most commonly used measure of body size, is
calculated by weight (kg) divided by height2 (m) 38' It is practical, easy to obtain,
reproducible and can be used in the clinical setting, as well as in epidemiologic
studies. In adult studies, correlations between densitometry estimate of body fat
composition and BMI unadjusted for age ranged from r2: 0.67-0.85.42 For
children, the correlation is lower, ranging from 0.55-0.77. There are limitations to
using BMI as a measure of body fat. It is a better indicator of heaviness because
it has difficulty distinguishing between that due to fat versus muscle. In children,
interpreting BMI is complicated by differing rates of growth and maturation. For
this reason obesity is defined relative to a selected percentile in a reference
group based on age and gender.

Childhood obesity is defined as a BMI above the 95th percentile relative to a
selected age— and gender-matched reference population, while overweight is
between the 85th and 95th percentile.” The 2000 us pediatric growth derived
from data from the NHANES studies (1963-1980) is used as the reference
population. A 3-4-unit change in IMI within a year is another measure of
excessive weight gain.

There is a chance for differential misclassification when self-reported measures
of weight and height are uSed to determine overweight status. This is because
overweight persons are more likely to under-report their weight thus introducing
differential misclassification. This effect would, however, bias the result towards

the null in the relationship between asthma and obesity.

The National Heart, Lung and Blood Institute and the World Health Organization
classified adult body mass index into categories (Table 1.2) based on deleterious
effects of BMI.

Table 1.2 Classification of Obesity

 

 

 

 

 

 

 

 

 

Adult Categories* BMI wt/ht2
Underweight <18.5

Normal 18.5-24.9
Overweight 25-29.9

Obese class 1 30-349

Obese class 2 35-399

Obese class 3 >= 40.0

Children : defined relative to age and sex
BMI >95th %tile Obesity
BMI >85%tile-95%tile Overweight

 

 

 

*(NHLBI and WHO)

Childhood BMI between the 85th and 95th percentile increases the risk for adult
obesity, while the likelihood of obesity related complications increase above the
95th percentile. A BMI >the 95th percentile in a child is analogous to an adult BMI
>30.

Asthma and Obesity: Similarities exist in the epidemiology of asthma and
obesity that may explain the recent interest in an association between the two
conditions. Both have risen in prevalence worldwide, especially in regions where
there is industrialization and urbanization. This rise has occurred during the same
timeframe. For example, pediatric obesity has doubled between NHANES l|
(1976-80) and NHANES III (1988-94),” while the prevalence of asthma
increased by 75%.2 Much of the rise affects similar populations, children,

minorities and low socioeconomic class.

10

Should a true causal association exist between asthma and obesity, the
population attributable risk (PAR) would be high. For instance, if we estimate a
2-3-fold increase in risk for asthma among obese persons and the prevalence of
obesity at approximately 25%, then the PAR will be 20-33 percent. The public
health impact of both conditions suggests that further investigation into the
relationship between asthma and obesity is warranted.

My thesis will examine the relationship between asthma and obesity. It will also
explore probable pathophysiologic mechanisms underlying this association,
specifically the relationship between body mass index and atopy using secondary

analysis of longitudinal data.

The following goals will be addressed:

1. Determine the association between BMI and asthma.

2. Explore the role of variables such as immunoglobulin E, eosinophil count
and breastfeeding.

3. Examine the effect of body mass index on FEV1, FVC, FEF50 and
FEV1/FVC.

4. Examine the role of body mass index at age 4 years, on measures of
atopy, pulmonary function and bronchial hyperresponsiveness ascertained

at 7 years.

11

CHAPTER 2

LITERATURE REVIEW ON OBESITY AND ASTHMA
Over the last forty years, there have been references to a possible association
between obesity and asthma?”23 These studies have not provided a clear
understanding of the nature of the relationship. Investigators disagree on whether
a causal relationship exists and have been unable to reach a consensus on
which is the risk factor or the outcome. One hypothesis proposes that reduced
energy expenditure in asthmatic children predisposes them to excessive weight
gain. The alternate hypothesis is that obesity is the risk factor in the development
of asthma. Supporting the alternate hypothesis are studies that show a greater
frequency and degree of bronchospasm in the smaller airways in obese non-
asthmatic children compared to their non-obese counterparts.”50
The association between asthma and obesity has been examined mainly in
observational studies because randomization of either disorder is neither feasible
nor ethical. In contrast to experimental study design, results from observational
studies are more likely to be plagued by methodologic issues of bias and
confounding. An example is case ascertainment of asthma. Reported history of
physician-diagnosed asthma or history of wheezing are the most commonly used
methods for defining asthma in epidemiologic studies. Responses to these
questions typically measure asthma prevalence and not incidence unless an
inception cohort, i.e, population with no prior known history of asthma or

wheezing is used. Even with inception cohort studies, where we assume we are

12

measuring incident cases of asthma, it is probable that some of subjects may be
misclassified if they have subclinical disease. Thus, most of the published studies
speak more to the association between obesity and severity of asthma rather
than the development of asthma. In addition, although BMI is used widely to
define obesity, its ability to accurately measure body fat versus heaviness is still
being debated.

The most challenging problem facing the causal link between asthma and
obesity, should one exist, is the multifactorial nature of both disorders; thus
teasing apart the role of the different factors in the relationship is difficult. It is
unlikely that a single identified risk factor acts independently, but rather complex
interactions between chest wall mechanics, genes, proinflammatory effects of
adipocytes and leptin, hormones, physical activity and diet underlie the
association. In addition, as we become more aware of the independent effects of
early childhood factors such as breastfeeding and birthweight on both asthma
and obesity, the role of these variables needs to be considered in the causal
pathway.

Tables 2.1, 2.2 and 2.3 summarize epidemiologic studies on the relationship
between asthma and obesity by study design. There is a preponderance of cross
sectional studies in published literature as shown in Table 2.1. Ten of the cross-
sectional studies found a positive association, while only one reported no
association. One of the eariier population based studies was carried out by Chen
et al.3 They analyzed the Canadian National Population Health Survey 1994-

1995. Asthma was determined by self-report of physician-diagnosed asthma.

13

Sample size was 17,605 subjects older than 12 years. BMI >=28 increased
prevalence of asthma in obese women but not in men. Stratifying the population
into two age groups, 12-24 years and above 25 years‘did not alter the results.
Their results suggest gender may be an effect modifier. A major concern for this
study is the validity of self-reported information. However, the possibility of
systematic bias introduced this way does not satisfactorily explain the observed
gender differences.

Another cross-sectional that found a gender difference in the obesity-asthma link
was a study of 9357 5- and 6-year old German children in two rural Bavarian
areas in Germany. 1° Obesity was characterized as BMI greater than the 97th
percentile. Children were defined as asthmatic if they had at least one physician
diagnosis of asthma, spastic or obstructive bronchitis. The severity of asthma
was assessed by number of wheezing episodes within the preceding 12 months.
In a multivariate logistic regression model, obese girls had a twofold risk of
having a diagnosis of asthma (OR 2.33; 95%CI 1.13-4.82). Other independent
variables in the final model included birthweight <25009rams, family history of
atopy, more than three febrile episodes in the first year of life and consumption of
whole milk three or more times a week.

Other studies have not found gender differences. Schwartz and von Mutius,
using data from National Health and Nutrition Examination Survey II (1976-1980)
and I” (1988-1994) respectively, reported a positive association in children

regardless of gender.5'7 Jang and Xu in adult studies report similar findings,

14

However, in Jang’s study, carried out in a high altitude rural area in Korea, the
risk of asthma was stronger in obese women compared with men."'"13
It is hypothesized that gender differences may be mediated by the effect of
estrogen and progesterone on lung function, especially estrogen. Supporting this
line of thinking is the change in gender dominance in asthma epidemiology from
male to female around pubertal period. Post-menopausal women receiving
hormone replacement therapy have been reported to have higher risks for
incident asthma.14 Also higher rates of skin reactivity to histamine corresponded
with peak estrogen levels in women tested at different stages of the menstrual
cycle.45 These mechanisms, although biologically plausible, still remain
speculative.

Castro Rodriguez et al.8 examined the association between asthma and obesity
in children aged 2 months to 12 years, using the NHANES III data. The study
population included very young children, which had not been previously studied.
In a multivariate model controlling for age, race, family history of asthma, and
environmental and smoke exposure, obesity was associated with asthma only in
African American children OR 1.64 (95% CI 1.2-2.3). Asthma was defined as a
history of physician diagnosed asthma. A major problem with case ascertainment
in this study is that preschool children are particularly prone to respiratory
infections with accompanying wheezing. Thus, diagnosis of asthma in this age

group is often difficult and imprecise.

15

Table 2.1: Epidemiological Studies on Causal Relationship: Cross-sectional

 

 

 

 

 

 

 

 

studies
Study Population Definition Definition Results
[sample of asthma of obesity
Chen3 N= 17605 Self-report BMI OR 1.74-1.84
Canada National population of for increasing
Health Survey physician BMI but only for
diagnosis females
Schachter‘ 17-73yrs Self-report BMI Asthma OR
Australia N=1971 of wheeze 2.04
white or Wheeze OR 2.6
3 pooled physician Medication use
epidemiologic diagnosis. OR 2.8
studies Histamine FVC, FEV1
inhalation (p<0.05)
Challenge-
F EV, FVC,
PEF, FEF
Schwartz5 Children Self-report 2 SD OR 1.6
USA 6months-11yrs of: a) increase in
5,672 physician skinfold in
Black 18% white diagnosis quintiles
82% b) wheeze
NHANES ll
Munoz6 14,908 History of BMI >90% OR 1.28 (95%Cl
UK 4-1 1yrs one Reference 1 .1 1-1.48)
white asthma group
representative attack in <10%
sample. last year or Skinfold
Inner city history of thickness
wheezing
von 4-17yrs Physician BMI OR 1.77 (95%Cl
Mutius7 NHANES III diagnosis 1.44-2.19)
USA of asthma
Castro N=12,388 Physician BMI OR 1.64 (95%Cl
Rodriquez8 2 months-16yrs diagnosis 1.20-2.26) AA
USA NHANES III of asthma only
Celedon9 N=7,109 Physician BMI BMI >30
China Anhui province diagnosis OR 2.3 (95%CI
of asthma 1.2-5.0)
Airway BMI <16 2.5
hyper- (95%Cl 1.4-3.8)
responsive
ness
Asthma
attacks

 

 

 

 

 

 

16

 

Table 2.1 (cont’d.)

 

 

 

 

 

 

 

 

 

von Kries‘" N=9357 Physician BMI OR for females
Germany 5-6yrs diagnosis, 2.33 (95%Cl
History of 1.13-4.82)
wheeze Males NS
Lusky'"I 17 years Self- BMI RR 1.8 (95%Cl
Israel N =12,000 reported 1.4-2.1)
history of for BMI <5%
asthma or 1.8 (95%Cl 1.4-
bronchitis 2.1) for BMI 5-
15%
NS BMl >85%
Jang12 Ages 50-93years Self— BMI Prevalence of
Korea High altitude rural reported wheeze higher
area history of with BMI >=25
N=707 wheezing (28.6% vs.
12.6%
p=0.01)
OR=1.1 (95%CI
1.03-1.18)
Sin Tr NHANES III (1988- Self- BMI Asthma OR
USA 1994) reported (quintiles) 1 .50 (95%CI
Ages >=17 years asthma 1.24-1.81)
N=16,171 FEV1/FVC Highest quintile
<80% had the lowest
risk of
FEV1/FVC
>80% (p=0.001)

 

The sole dissenting cross-sectional study was conducted in Israel among 17-

year-old army recruits". (Table 2.1) Standardized questionnaires and

anthropometry measurements were obtained at screening health programs in 5

induction centers. They used BMI and characterized each recruit according to

his specific ethnic group distribution. Ethnicity was defined as country of origin of

the paternal line. The authors found a RR of 1.8 (95%Cl 1.4-2.1) for severely

undenlveight boys (BMI <5%tile) and 1.8 (95% CI 1.4-2.1) for mild underweight

(BMI 5-15%tile), but no increased risk with BMI >85%. The greatest threat to

17

 

internal validity in this study was the imprecise definition of asthma. Asthma was
grouped with “chronic bronchitis” in the analysis. The study population was only
males, thereby limiting generalizability of results. The findings in this study, if
true, refute the hypotheses that asthma leads to sedentary behavior and then
obesity.

Interestingly the study by Sin et al. using NHANES III data, found contradictory
results.13 Even though physician diagnosis of asthma was significantly
associated with obesity, children in the highest quintile were least likely to have
an FEV1/FVC less than 80%.

Their result suggests a separate mechanism, not mediated by decreases in end
tidal volume with resultant bronchospasm. It is possible that obese children utilize
more health care services, with greater opportunities to be accurately or
inaccurately diagnosed as asthma. Indeed a study of clinic practices in Michigan
found higher health care utilization for obese children compared with their normal
weight peers. ‘6

The odds ratio reported by the cross-sectional studies (Table 2.1) range between
1.28-2.6, reflecting a weak association between asthma and obesity. It is
probably that association is stronger, but misclassification of asthma, or obesity,
errors in data collection or report can bias the results. Additionally by their
inherent design, these studies provide limited evidence of causation.

Table 2.2 contains cohort studies, which provide the strongest evidence to date
to support the premise that obesity is a risk factor for asthma. There have been

few cohort studies on the obesity-asthma link in the literature. In a prospective

18

cohort study of 85,911 female registered nurses, ages 24 to 44 living in one of 14
us states“, high BMI was positively associated with the onset of physician-
diagnosed asthma. The study population was part of the Nurses Health Study.
Baseline information was obtained in 1989 with subsequent follow-up
questionnaires in 1991, 1993, and 1995. There was greater than 90% response
rate for all the surveys. Diagnosis of asthma was defined by: (1) reiterated
physician-diagnosis on supplementary questionnaires and use of asthma
medication since diagnosis; (2) above, p_lu_s prescribed controller medication in
past year (e.g., inhaled corticosteroids); and (3) above, p_lu_s physician-diagnosis
within “one month” of symptom onset. The specificity increased with later
definitions. Exposure was measured using BMI and waist/hip ratio. Validation
studies within the cohort on self-reported weights and diagnosis of asthma
showed excellent correlation between self-report and actual measures. In a
multivariate logistic regression, controlling for 9 factors: age, race, US region,
smoking status, physical activity, total energy intake, hysterectomy status,
birthweight, and duration of breastfeeding, the relative risk of adult onset asthma
was strongly associated with an increase in BMI (p< 0.001) with all three case
definitions. Using their strictest case definition, weight gain greater than 25kg
since age 18 years, among the subgroup, had the highest risk 4.7 (95%Cl 3.1-
7.0) The BMI-asthma association was found in every subgroup examined,
including those with a recent health screening exam, those who used nutritional

supplements, nonsmokers, past smokers, and current smokers.

19

Table 2.2: Epidemiological Studies on Causal Relationship between
Asthma and Obesity: Cohort studies

 

 

 

 

 

 

 

 

 

 

Study Population Exposure Outcome Results
I sample ‘
Camargo” Nurses health BMI Questionnaire RR 1.1-2.7 for
US study. WHR asthma by increasing BMI
no. 85911 wt @18yrs report (p<.001)
ages 24-44 3 case for trend
inception cohort definitions
Shaheen15 8,960 BMI >30- Self-report of 1.72 (95% CI
UK 26 years obese history of 1.29 to 2.29).
22.5-29.9 — asthma and/or Higher risk for
overweight wheeze. females
Positive history
at age 10 by
maternal self-
report or
physical exam
Chen"; 20-64yrs BMI Self-report of OR females
Canada 9149 physician 1.9 (95%Cl 1.1-
diagnosis of 3.4)
asthma males OR 1.1
(95%Cl 0.3-3.6)
Xui'7 Finish birth BMI Self-report of ‘ BMI >=95%tile
Finland cohort Physician in adolescence
1966-1997 diagnosis of OR 2.09 (95%Cl
F/u at age 31 asthma 1.23-3.57)
years Young
N: 4719 adulthood
OR

 

The study used only incident cases, thus showing a temporal relationship

between obesity and clinically apparent asthma. However, because no objective

assessment of pulmonary function was done at the beginning of the cohort, it is

possible that subclinical disease existed. Also the spectrum of asthma across life

spans differ widely, thus it is conceivable that the women with a new diagnosis of

asthma had a prior history of resolved asthma in childhood. Other limitations of

the study include poor generalizability because of the make-up of the cohort and

20

 

use of self-reported measures. Although the use of a knowledgeable cohort may
provide more accurate reporting of diagnosis, there is a possibility of over-
reporting of illness or the cohort may be comprised of-a group of individuals with
healthier lifestyle behaviors than the general population. The latter occurred in
this cohort. The mean weight of participants in the Nurses Health Survey was 3.3
kg less than their counterparts in the US population.

Shaheen’s study, using a historical British cohort, showed a positive relation
between increasing BMI and asthma when both exposure and outcome was
ascertained at age 26 years.15 The participants completed a survey at age 26
years. The prevalence of asthma rose with increasing adult BMI. After controlling
for birth weight and other confounders, the odds ratio comparing highest with
lowest quintile of body mass index was 1.72 (95% CI 1.29 to 2.29). However,
BMI at age 10 years was unrelated to a diagnosis of asthma at the time of the
survey (age 26 years), but low birthweight (<2kg) was a risk factor for asthma
(OR 1.99, 95% Cl 0.96 to 4.12). Although this study reported a positive
association between asthma and obesity, it did not demonstrate that obesity
preceded development of asthma.

Chen et al. reported a positive association between obesity and asthma, but only
in females, in a population-based study.16 They used data from the first and
second cycles of the National Population Health Survey, conducted in Canada in
1994-1995 and 1996-1997, respectively. Over the tvvo-year follow up period,
2.2% of the participants developed asthma, with twice as many women

compared to men (141 vs. 68). The gender difference observed may reflect the

21

small number of men in the study with the outcome rather than a true underlying
biological reason.

The cohort studies provide evidence of temporality, a strong criterion for
causality. Three of the studies report a positive association, albeit their findings
were limited to women in two of these studies. The third study found a stronger
association in women compared to men. The biologic plausibility of a gender
difference in an obesity-asthma link is still poorly understood. Thus, it is entirely
possible that an unmeasured confounder that is strongly related to females may
account for the association reported. With only four studies with differing results
published to date, further studies in different populations and better
ascertainment of confounders are needed to substantiate a relationship.

All the case control studies were clinic based except for the study by Belmarich
et al.22 which had patients from the emergency department as well as primary

I.18 and Luder19 examined the hypotheses in the

care clinics. Gennuso eta
pediatric age group, using case-control study design (Table 2.3). Both studies
recruited patients from inner-city clinic populations and the racial make-up was
predominantly Hispanic and African American. Obesity was defined as BMI
>=95th percentile for age and gender and overweight above the 85th percentile
and BMI <85th percentile as the referent group.

Gennuso's study found there were significantly more obese children with asthma.

This finding was more marked in the group above the 95‘“ percentile for BMI (RR
2.6

in obese children vs. 1.9 in overweight group). Asthma severity, however, was

22

unrelated to obesity. Luder's19 study was conducted in New York (East Harlem

and south Bronx neighborhoods). Cases were drawn from a tertiary

pulmonclogy clinic. The controls were selected through a stratified sampling

design of schools in all of New York City.

Table 2.3: Epidemiological Studies on Causal Relationship between
Asthma and Obesity: Case-Control studies

 

 

 

 

 

 

 

Study Population I Exposure Outcome Results
sample
Gennuso 176 controls Asthma BMI RR 1.9 -2.6
‘8 86 asthma divided by No relation with
USA 4-16 yrs severity severity or steroid
Medical use
record. review
Luder19 2-18 yr old Asthma BMI OR 2.05-6.34 with
USA black and classified different indicators of
Hispanic by severity severity
children. 1.3 RR (95%CI 0.99-
Controls 6-13 1.52, p=. 06)
yr. old in same BMI>=85th percentile
school district 1.5 (95%CI 1.05-
Medical record 2.19, p=. 03)
review. > 95th percentile.
Young20 17-96 yrs Physician BMI OR, 1.2; (95% Cl,
USA 386 cases diagnosis 25—29.9 1.1-1.4);
744 cases use of BMI 35-40 OR, 4.8; (95% CI,
bronchodi- 2.6-9.1 )
lator
Brenner“ 265 cases Physician BMI Prevalence of
USA 482 controls diagnosis overweight in
AA asthmatics (16%)
non asthmatics
(15%)
Belmaric Children Skin BMI >95tfi Increased severity by
h22 1-4 yrs testing Reference 5- use of steroids
USA 1,528 PEFR 95‘"%tile ER visits in obese
9 month f/u Steroid use asthmatics
AA-64% ER
Hispanic-28% utilization

 

 

 

 

 

23

 

 

 

Guerra23 Nested case Physician BMI OR of BMI >28
USA control diagnosis assessed 2.1 (95% Cl 1.31-
102 of incident before onset 3.36)
asthmatics asthma asthma Association stronger
age >= 20 ‘ for women OR 3.45
years BMI 25—27.9 (95%Cl 2.10-5.67)
BMI >= 28

 

 

 

 

 

 

This was probably done for convenience because they used the data from the
New York State Department of Health. Both controls and cases had different
person year experiences in relation to asthma thereby violating the study base
principle. Additionally, the controls were 6-13 years old and the cases 2-18 years
old. The authors acknowledged that some of their controls might have been
overweight thus introducing misclassification. They found a 1.3 RR (95%CI 0.99-
1.52, p=. 06) for asthma among those with BMI>=85th percentile and 1.5 (95%Cl
1.05-2.19, p=. 03) for the 95th percentile. Ovenlveight status was related to certain
measures of severity, e.g., use of more than 3 asthma medications (OR 2.05
95%CI 1.02-4.13) and peak expiratory flow rate =< 60% of predicted (OR 6.34
95%Cl 1.83-22.0). Both studies have limited generalizability to other populations.
The only case-control study that reported a negative association between asthma

l.21 in a sample of predominantly African

and obesity was by Brenner eta
American children. The prevalence of obesity among asthmatics and non-
asthmatics was similar (16% vs. 15%).

Inherent biases in the case control study design, particularly with case
ascertainment, undermine the significance of the weak positive association found

in these studies. It is unlikely that further case control studies examining only the

association between both disorders will be especially helpful to the current body

24

of literature. Studying the role of potential mediators or confounders in the
obesity-asthma link using this study design may be more useful.

In Table 2.4 are studies that examined the causal relationship from a different
perspective. Their proposition is that if weight is the risk factor, then reducing or
increasing the weight should affect the outcome. This approach strongly supports
the causal criterion of specificity. Although these studies do not necessarily
address incident asthma, they provide evidence that weight is directly related to
asthma symptoms. Hence, investigating pathophysiologic mechanisms that are
common to both conditions can help elucidate the causal pathway.

Chen et al.48 in the Humboldt cohort study (1977-1983) in Canada, examined the
effect of weight gain on lung function parameters. Their results Showed that
mean FEV1 and FVC were highest in the group that gained < 1.0kg and lowest in
the group that gained >=4.0kg (p.001 ). Chen's study controlled for age, gender
and race, and smoking status (Table 2.4). Interestingly when the mid maximal
expiratory flow rate, which is sensitive for mild impairments in lung function was
used, the result was not significant.

Stenius-Aarniala et al.47 and Dixon et al. 5‘ hypothesized that decreasing body
mass index should lead to improvements or resolution of asthma symptoms.
Stenius-Araniala conducted a randomized controlled trial in Finland. Participants
were recruited by advertisements in the newspaper. They had a history of
physician-diagnosed asthma, were ages 18-60 and had a BMI of 30-42.
Intervention (n=19) received weight reduction (dietary therapy for 8 weeks) and

participated in a discussion group, while the controls (n=19) only took part in the

25

discussion group. The outcome for the study was improvement in pulmonary
function tests and asthma symptoms. The weight loss was 14.5% in the
intervention group and 0.3% in controls. There were statistically significant
improvements in FEV1, FVC, reported shortness of breath and use of rescue
medications and number of asthma exacerbation. A limitation in the study was
lack of blinding to the treatment, which could introduce biases. But more
importantly, the authors failed to adjust for the baseline differences in pulmonary
function test (PFT) in both groups, a measure of lung function that is closely
related to symptoms of asthma and medication use.

Dixon’s study evaluated the prevalence of asthma in the morbidly obese and

the changes in asthma after Iaparoscopic adjustable gastric banding (LAGB)
surgery for morbid obesity.51 This was the first study on weight loss with asthma
as the primary outcome. Asthma was assessed preoperatively in all patients
presenting for LAGB. 32 consecutive asthmatic patients were followed up
clinically and by a standard questionnaire for at least 12 months after surgery,
and any change in asthma impact determined as an asthma score was recorded.
The asthma score included severity, daily impact, medications needed,
hospitalization, sleep, and exercise. Neither the physician nor patient was blinded

to the outcome of interest.

26

 

 

 

 

 

 

 

 

 

 

 

 

.mcmmo 92mg mm. E» 3.0
5 8:9ch EmoEcQw c.0550
.wcmmo cacao
ozmepzmmcoc cmcno mm“. .5: 3 >36
5 30 .v3 coco—Ewan .mnmn .>wn_ mmcconu mommo 65:8 <w3
mac 5 :9. $9 came—5:830 2255 cacao me memo Bamamx
exec o>n_:>mn.
cacao E o>n .Emn. 0?. .EE _ wmz<zz a; 9-:
c_ 838% $8555 Emmamococofi doc mate: co .989 65:8 33m
.8me magma 3885 case Sow co m: can 63:8 BEE.
mm. 5 cocccoEu oz 0&9me scour. =2m momma em cwmo e4.89.0.0
cam Ego; 9. en o>nE>mn 2“. as.» a
5:5 689m. 0?... new 0>n. 5mm 2mm» mmrmw >36 mnmcmo
Emu. .828. caps. P>mn as; £925 938 Be :28 885 acoaazoeao
ad? a 9E: a; 8.?
save v 9.; e O an. .Ecaoa «Tom Sm Egg...
.3 new mxsw© .cE co c2838. 8.02:8 A
be. cc Ease Ba 0?. Ease .e>mn_ Ego; 2 can essence. 852e,
.F>mn_ E 2399?. 6?. .Emo 8.83 m cacao 2 5m cease _oEoE_cocxm
2:39.:
3:501 choSO cozcoEcE. cows—anon on»... 255

 

 

moo. EH33 no “come 85 9.3.888 353m "95:03.23. .330 cc cofiam ficfiSSEcuEm SN Ens...

27

 

 

 

 

we? .m><
8; :-~ 2“.
>E2moca._c_ 33m
3985 5328:. >938“. ov :o_Ec>
0.3.6 “—0 £9.95 no em: 53> 833 .55 emu—2 ESE <w3
9636 9:53 co 355m 2.5mm m; no-8 c>z 85
55:02» 003280 mE£w< .55 £394 ov .8391 @9ng
as mm... choEm
293839 .cmmcflfico 9.2, m cm 3E .52 >36
.mod v a new .mood cEEmEI 9.3 m x mosz5m< co_Ec>
v a .83 n 3 mac. 6?. 8:52.25 E .93 does
2995 5cm communes €>wn .nwn >565 m; 3.3. mg “ESE“.
0?. can .A :5“. .nmc essence to Sm 9.22 ceased aces:
3:93 >893
958:2 9.6ch moszfimm >36
$80!: Boom Boom oEmmm cacao co_Ec>
mEfimm E concave. mE£m< .._>F© cfifimsfim 2299: ESE c>= czmbmze.
Eamon EmoEc9m 95502330 93 mm acumen. 5:050

 

 

 

 

 

 

 

 

3.233 «N 2%..

28

Participants had a mean body weight of 125.2 kg and a BMI of 45.7 kg/m2
prior to operation, and a weight of 89.3 kg (BMI 32.9 kg/m2) at follow-up.
The mean preoperative scaled asthma score was 44.5 +/- 16. There was a
highly significant reduction at follow-up to a mean value of 14.3 +/- 11 (P
<0.001).

In both studies, mechanisms other than direct weight loss may be responsible
in the improvement noted (i.e., gastroesophageal reflux). More challenging is
teasing apart less tangible factors such as the effect of improved emotional
state and self-esteem on perception of symptoms. This is important because
even though overall symptoms of asthma improved, use of rescue medication
was unchanged in Araniala's study suggesting that ainlvay hyperreactivity
persisted.

Kaplan‘gexamined the relationship between exercise-induced airway
reactivity and body fat among non-asthmatic obese children using non-obese
controls (Table 2.4). Cases and controls were recruited from a medical care
center. Exclusion criteria included history of wheezing, asthma, bronchitis,
bronchodilator use and cardiopulmonarydisease. Body mass index and
skinfold thickness was exposure of interest. Pulmonary function test with
exercise challenge was done. They found obese subjects were more likely to
have exercise-induced asthma versus the controls (69% vs. 42%). Also
there was a significant correlation between a 15% decline in pulmonary

function parameters and obesity.

29

Gokbel conducted a similar study with Turkish children.50 BMI percentiles
from the NHANES I study were used to define obesity. He also found
Significant differences between obese and non-obese subjects in FVC and
PEFR, but in contrast to Kaplan's study found no difference in exercise-
induced bronchospasm in both cases and controls. The average change in
the parameters among the obese children at 5 and 10 minutes was 3% and
7% respectively.

Although both studies found significant differences in obstructive lung function
measures between obese and non-obese children, the differences are small
and will be unlikely to translate to a clinically significant asthma on an
individual basis.

Summary: Most studies to date have reported a positive association between
asthma and obesity. However, a causal association between asthma and
obesity remains tenuous as components of Sir Bradford Hills causal criteria
are lacking or have not been adequately substantiated. The strongest
evidence supporting causality has been temporality provided by three cohort
studies, but two were limited to women thereby hampering external
generalizability. Muddling the issue of temporality is the shifting clinical
spectrum of asthma over the life stages. It is yet to be teased out whether
resolved asthma in childhood is a separate entity from new onset adult
asthma. Strength of association ranges between risks of 1.0-2.5 in most
studies. These values do not represent a convincingly strong association

between both disorders. Varied definitions for obesity, particularly in adult

30

studies, weaken the consistency of a positive association across studies. The
validity of the positive relationship is further hindered by the preponderance of
observational studies, with their inherent methodologic weakness, published
in its support. Because most studies used prevalent asthma as the outcome
or exposure, they provide evidence that obesity is associated with asthma or
increases its severity, not necessarily causality.

The multifactorial nature of both disorders and our lack of understanding of
the role of third variables such as breastfeeding, genes, and hormones make
unidentified confounding a real possibility. Lack of breastfeeding for instance
has been independently associated with both asthma and obesity. Failure to
control for this factor can lead to a spurious association.

A positive association has been reported in different populations, age groups
and times supporting consistency of results. The dissenting studies are fewer
and limited by case ascertainment biases. However, dose response has been
inconsistent. Only five studies report increasing risk or severity of asthma with
higher levels of BMI.3'1“'19"""52 Studies that report weight gain or weight loss
correlate with the risk of asthma provide evidence of specificity and suggest
that indeed weight is the problem.” "'53 These studies, however, do not
directly provide evidence of causality in the development of asthma. Evidence
supporting other causal criteria of biologic plausibility, analogy, coherence
and experimentation are sparse and remain largely speculative. Finally, as

with all reviews limited to published studies, publication bias poses a real

31

threat as studies reporting a positive association are more likely to be
published.

In conclusion, the possible confounding of genetic, pathophysiologic and
environmental factors is probably the most significant threat to a causal
relationship between asthma and obesity. Most probably, a linear relationship
between asthma and obesity does not exist. The possible confounding of
genetic, pathophysiologic and environmental factors is probably the most
significant threat to a causal relationship between asthma and obesity. This is
further complicated by probable interactions between risk factors.

Based on existing literature, we can deduce that some relationship exists
between asthma and obesity. I have thus developed 5 models that critically
explore the different ways both conditions could be associated. These models
are not mutually exclusive and allow us to speculate on the role ofldifferent
factors drawing from collaborating evidence in the extant literature where
available. Some models are better developed than others. Limiting our ability
to make causal conclusions are the paucity of cohort studies or experimental
studies that examined incident asthma. The key objective of developing the
models is to consider all possible ramifications of an association, as there is
inadequate evidence based on the existing literature to prove the models.
Understanding or refuting each model requires a better understanding of the
role various variables can play in the obesity-asthma link. Table 2.5
summarizes the hypothesized role of different variables, which may be

mediating, confounding or antecedent in the causal mechanism. A mediator

32

or intervening variable acts as a true causal pathway, while confounders tend
to obscure the true relationship, thereby suggesting spurious associations; If
A leads to B which in turn causes C, B is termed a mediator or intervening

variable. Variations in C can be caused by B, which in turn can vary based on

levels of A. For example, obesity causes gastroesophageal reflux (GERD)

Table 2.5 Proposed role of third variables in the relationship between

 

 

asthma and obesity

Antecedents Mediators or Confounders
Hormones, e.g., Hypothalamic- Hormones, e.g, Estrogen
Pituary- Adrenal axis Gastroesophageal reflux
Underlying genetic or familial Mediators, e.g., leptin, COX-2,
disposition lnterleukins
B-adrenergic receptor responsiveness Atopy
Level of physical activity Diet

Body fat distribution

Level of physical activity

 

 

 

 

which leads to asthma. On the other hand, a confounder is a true cause of
the outcome. It is associated but not caused by the exposure. Using the
earlier example, GERD could be merely associated, not caused by obesity,
but GERD is a risk factor for asthma. In this scenario, GERD functions as a
confounder. Thus, the same variable can potentially be a confounder or a
mediator of the obesity-asthma link. Mediating and confounding variables are
difficult to distinguish on statistical grounds alone. Controlling for any of these
variables, if they are mediators, will produce an erroneous nonsignificant

relationship between asthma and obesity. On the other hand, failure to control

33

for them, should they be confounders, will lead to a type 1 error, leading us to
assume a positive association when none exists.

An exposure that precedes a mediator in the causal chain is an antecedent
variable. Some authors refer to this as a simple antecedent. An antecedent
variable (A) usually occurs early, before both variables are noted to be
associated ( B is associated with C). An example would be if obesity causes
asthma, and obesity itself is caused by lack of breastfeeding, then the
antecedent variable in this relationship is lack of breastfeeding. A simple
antecedent is not a confounder, because it's associated only with the
exposure. However in complex associations, an antecedent can act as a
confounder. Consider the example on lack of breastfeeding described earlier.
Lack of breastfeeding, which is considered the exposure, confounds the effect
of obesity on asthma, if lack of breastfeeding contributes to asthma by other
mechanisms (e.g., stimulates Th-2 helper cell dominance over Th1 cell type).
Finally, both extremes of BMI have been linked to asthma suggesting a
probable J-shaped relationship between asthma and BMI.“'11 A J-shaped
curve has been reported for BMI and other health outcomes, such as
mortality, suggesting this may represent a valid model.“"“ In this instance,
BMI acts like an effect modifier. Various levels of BMI exert different effects
on the exposure. What remains unclear from the literature is the exposure
BMI exerts its effect on. Decreased levels of physical activity or specific
components of dietary intake common to both undenrveight and ovenlveight

individuals are plausible exposures. It is hoped that information gleaned from

34

reviewing the models will direct further studies and statistical modeling
techniques that will delineate the true association between the two variables.
Conceptual models of causation: In Model 1 obesity-is a risk factor for
asthma, mediated by poor chest wall mechanics or through cellular
mechanisms. Alternatively, in Model 2, asthma may indeed be a risk factor
for obesity. Model 3 posits that the association is spurious. The association
occurs as a result of a variable that is common to both disorders, e.g., low
birthweight, genes. BMI is an effect modifier in Model 4. Different levels of
BMI have varied effects on development of asthma. Finally, Model 5,
suggests there is no relationship between asthma and obesity at an individual

level, only at a group level.

Model 1: Obesity Causes Asthma

a) Chest Wall
mechanics

- b) Atopy e.g. lgE
Obesrty _> c) lnfammation-lL-G, —> Asthma

VLDL,
d) Gastroesophageal
reflux

In Model 1, obesity is a risk factor for asthma. Postulated mechanisms
include altered chest mechanisms, atopy, immunological pathways,
hormones, gastroesophageal reflux and diet.

Pulmonary Function and Chest Wall mechanisms (a): At the physiologic
level, a fall in lung capacity, with resultant small tidal volumes, leads to the
development of slow cycling acting-myosin cross bridges in bronchial smooth

muscle referred to as the “latch state”. The latch state has been proposed as

35

the reason for sustained airway obstruction in hyperactive ainrvays.
Dysfunctional chest wall mechanisms due to extrinsic chest wall and intra-
abdominal adipose tissue is hypothesized to prevent full downward excursion
during deep inspiration with a decrease in FEV 1, FVC, total lung capacity.
This in turn, can result in development of the latch phase with increased
bronchial muscle tone and hyperresponsiveness.

A greater preponderance of studies report a negative correlation with FEV1,
FVC, FVC/FEV1 and MMFR with different measures of obesity in children
and adults.“57
Collins et al.54 in a study limited to 42 normal to mildly obese Caucasian '
firefighters, found upper body fat distribution was associated with modest
impairments in FEV1, FVC and TLC. This study suggested that upper body
fat impedes ventilatory excursions leading to impaired lung function. Major
shortcomings in this study are its lack of generalizability and control for
smoking, a potential confounder as well as the use of Skinfold thickness, a
notoriously technician-dependent measure. However, the investigators
addressed this potential source of bias by using a single observer and taking
the average of repeated measures.

The Normative Aging study, a longitudinal study of 2280 men 21-80 years
without chronic illness, also found an inverse association between FEV,
FEV/FVC ratio and increased BMI and body fat distribution but only among
50-59 year-old men (p<0.01).55 The study did not control for smoking.

Differential mortality rates leading to survivor bias, e.g., expected higher

36

mortality rates among older men and obese participants, does not adequately
explain their findings as one would expect greater mortality rates among the
groups older than 59 years.

The study by Ray et al.56 did not find abnormalities in obstructive ainrvay
function. They examined pulmonary function in 43 healthy nonsmoking,
obese adults (mean age 25 years) using spirometry, lung volume
measurement by nitrogen washout and Single breath diffusing capacity for
carbon monoxide. All subjects exceeded 153% of predicted body weight.
They had no control group but compared their results with standardized age
predicted values. Analysis was done in groups of increasing obesity. They
found measures of inspiratory capacity such as Tidal Volume, Total Lung
Capacity and Vital Capacity were within normal predicted values for age and
sex. The finding of normal inspiratory capacity, regardless of the severity of
obesity, indicates normal lung compliance without intrinsic disease. Thus,
pulmonary function abnormalities in obese subjects may be secondary to
extrinsic lung disease independent of obesity.

In conclusion, it is interesting to note that despite the decrease in predicted
values in these studies, most of the subjects remain within the 95%
confidence intervals of the predicted population values, casting doubts that by
this mechanism alone, these subjects would present with clinically apparent
asthma.

Another proposed mediator in Model 1 is atopy. (b) Atopy: Atopy refers to

allergic conditions that tend to cluster in families. The phrase atopy, which

37

means “strange disease", was coined as early as 1923 by Coca, who
observed that in a group of people with manifestations of rhinitis or asthma,
no antibody could be detected in vitro. With improvements in scientific
technology, the pathophysiology of atopy involving immunoglobulin E has
been better understood.

The prevalence of atopy is difficult to accurately estimate, as the definition
remains controversial. Most studies use a combination of presence of ‘
symptoms and elevated lgE or Skin reactivity. It is estimated 20-30% of the
population are atopic; 50% of them are children.58 There is a 2-3 fold risk of
atopy in the offspring if both parents are atopic, while males are Slightly more
likely than females to have atopy.58 Along with the rise in asthma, a parallel
increase in the prevalence in allergy and atopy has also occurred.
Urbanization, greater exposures to indoorallergens and reduction- in the
practice of breast feeding have been proposed to explain the rise.

Allergy sensitization in asthmatics occurs for both indoor and outdoor
allergens but is more pronounced for indoor allergens. Some studies have
reported odds ratios as high as 19.7 for dust mites.59 Obese individuals are
more sedentary and thus are likely to spend more time indoors thereby
increasing sensitization to allergens in susceptible individuals.

There is a strong association between total and specific immunoglobulin E
and asthma.58 T cells, particularly Th2 type, primarily initiate physiologic
responses in atopy. These cells stimulate inflammatory response in the

presence of the offending antigen by releasing lL-4 and lL-5, which lead to

38

lgE production, eosinophilia and mast cell degranulation. The
proinflammatory mediators of histamine, tryptase, prostaglandins, and
leukotrienes result in the immediate and delayed reactions seen during an
asthma attack. In addition, there is an increase in lung epithelial cell adhesion
molecules for basophiles, eosinophils, lymphocytes and vascular
permeability. As these cycles continue to unfold, the histopathologic signs of
chronic lung inflammation take place leading to airway hyperresponsiveness.
The effect of BMI on atopy has not been extensively studied.”61 Huang et
al.60 in Taiwan recruited 2028 eighth grade children from seven junior high
schools, one in Taipei City, three in towns and three from rural areas, in a
cross sectional study examining the effect of BMI on allergy. All the students
in randomly sampled classes completed a questionnaire, which was modified
from the ISAAC core questionnaire. To minimize misclassification bias, BMI,
atopy and BHR were derived from objective measurements. Atopy was
defined by a skin-prick test using six allergens. The study population was
grouped into quintiles of BMI by sex. In a multivariate analysis adjusted for
area of living, sibling number, parent education level and family history of
asthma, using the middle three quintiles, as references, girls in the highest
quintile had increased risk of atopy. (Odds ratio=1.77, 95% confidence
interval=1.15-2.73).

A cross-sectional field health survey with 1129 preadolescent children in
Poland concurred with Huang et al. finding a positive association between

BMI and allergy.61

39

Other studies have found no effect.7 In a recent study on 7505 children aged
4 -17 years, von Mutius et al. found a positive association between asthma
and obesity but no relationship with hay fever, atopy or eosinophilia after
adjusting for age, sex, ethnicity, household Size, study area and passive
smoke exposure.7 Atopic sensitization was determined by skin prick tests.
Across quartiles of BMI, using Z-scores, the prevalence of atopy was similar
ranging for 48.6% in the 1st quartiles to 53% in the 3rd and 4th quartiles. The
effect of BMI on asthma remained the same, when atopy was added to the
model. Their findings suggest it is unlikely atopy is a mediator in the obesity-
asthma link.

The role of atopy, if any, in the relationship between obesity and asthma
remains inconclusive. The Mutius study, which found no association, was the
only study to use objective assessments of atopy, thereby strengthening the
credibility of its results. Further studies are needed to explore the role of
atopy, especially since allergy mediated asthma is a dominant form of
asthma.

Inflammatory Mediators: Inflammatory mediations and markers have been
linked to obesity and asthma and may serve as mediators (Model 1).
Expression of interleukin-6, a proinfiammatory cytokine in adipocytes,
production of cycIo-oxygenase 2 (COX-2) with stimulation of prostaglandin,
and beta-adrenergic receptor reactivity to these stimuli, amongst others,

suggests a pathophysiologic basis that may explain the association.

4O

Adipose tissue expresses TNF and lL-6, which are potent inflammatory

mediators. IL-6 has been associated with IL-4, lL-2, and TNF stimulation and

cytokines implicated in asthma. In addition, leptin, which is a protein,

secreted by fat cells, has been associated with interleukins. Figure 2.1

illustrates pathophysiology of airway inflammation and the effect of adipocytes

andlepfin.

Figure 2.1 Schematic diagram showing pathophysiology of airway

inflammation and effect of adipocytes and leptin

Allergen
+
Specific lgE

 

 

BMI

 

 

Airway
Inflammation and
bronchial
hyperresponsiveness

 

 

 

   

i; Adipocytee
Leptin

A group of investigators tested the hypotheses that overweight or obesity is

associated with low-grade inflammation measurable by an acute phase

reactant, the C-reactive protein (CRP). 25 The study population comprised of

41

non-institutionalized civilians, 7938 men and 8678 women, aged 17-39 years,
from the third National Health and Nutrition Examination Survey (1988-1994).
Seventy-six percent were Caucasian and 34% smokers. Analysis was
adjusted for potential confounders like smoking, cardiovascular disease, and
inflammatory disease. They reported an OR 2.13 (95%CI1.56-2.91) for men
and 6.21 (95%CI 4.94-7.81) in women for increased C-reactive protein in
obese subjects. The rise in CRP was directly related with increases in BMI.
Restricting the analyses to young adults aged 17-39 years did not alter the
findings. A limitation in interpreting the results is the use of a single
measurement of CRP. This may indicate active inflammation at a certain point
in time rather than an on-going process. Although this study was carried out
in the context of cardiovascular disease, the process of inflammation in
atherosclerosis is analogous to that seen in asthma, so this mechanism
presents a viable explanation for the obesity-asthma link.

I.26 investigated the release of histamine, a proinflammatory

Gonen eta
mediator, from circulating basophils. Basophils were extracted from a blood
sample taken from 28 healthy volunteers and incubated with very low density
lipoproteins (VLDL). Because VLDL are associated with obesity, it is
conceivable that the bronchospastic and inflammatory findings in asthma may
be partly mediated by this mechanism. The cells from each subject were
analyzed independently. Histamine release following incubation of basophils

with lgE was used as the control. VLDL showed similar kinetics in cellular

response compared with lgE, a well-known immunologic factor in the

42

pathogenesis of asthma. Incubation of basophils with VLDL showed a dose-
dependent cellular histamine release with peak histamine release occurring at
about normal body temperature. Fifty percent of the maximal histamine
release occurred after 5-12 minutes for both VLDL and lgE. Cyclic-AMP
agonists known to inhibit lgE induced histamine release, also showed
Significant inhibition of VLDL-induced histamine release. There was a lack of
correlation between the lgE and VLDL response among different donor
basophils, suggesting different mechanisms of action.

This study, while not directly related to an asthma-obesity link, sheds some
light on possible pathophysiologic mechanisms. The accuracy of invitro tests
in elucidating cellular mechanism such as those used in the above study has
been repeatedly demonstrated; however, in vitro tests may not accurately
reflect the process in vivo.

Inflammation may represent an entirely valid mechanism, but studies linking
the inflammatory process to obesity and airway hyperactivity are needed
before a causal relation can be assumed.

Gastroesophageal reflux (GER): Gastroesophageal reflux has been
strongly associated with asthma and obesity in multiple studies. Thus, it may
explain the association between both conditions (Model 1). It is still unclear
whether GER is a causal agent or an exacerbating factor in asthma. The
prevalence of GER in asthmatics ranges from 34-89% in studies of both adult
and pediatric population”. Andze et al. identified severe reflux in 41% of

children with asthma.63

43

Underlying pathophysiologic mechanisms include: 1) chest expansion with
diaphragmatic flattening leads to air trapping and negative intrathoracic
pressure. Against a positive intraabdominal pressure, the integrity of the lower
sphincter becomes impaired allowing for reflux of gastric contents into the
esophagus; and 2) gastric acid within the esophagus initiates a vagal
response causing a bronchospastic reflex within the airways.

Reflux in the lower esophagus is also thought to potentiate other
bronchospastic triggers. For instance, a provocative dose of methacholine
instillation that provoked a 20% reduction in forced expiratory volume in one
second was much smaller following esophageal acid exposure.64 Other
mechanisms include microaspiration of gastric contents, medication used in
treating asthma such as xanthine derivatives, and systemic Beta 2 adrenergic
agonists, which lower LES tone.

The association between obesity and GER is less well corroborated with
scientific evidence. While some studies demonstrate an increased rate of
GER with obesity, others refute this.6569 The strongest evidence that both are
connected stems from the improvements noted in GER symptoms following
weight loss.65

Establishing a temporal relationship between obesity and asthma is key in
Model 1. Few studies are prospective (Table 2.2). Although obesity
precedes asthma in these studies, we have to be cautious in the
interpretation of “incident asthma”, particularly in the adult population. None of

the cohort studies have controlled for wheezing in childhood. Children with

44

transient wheezing in childhood, or children who outgrow their asthma as
adolescents, can develop asthmatic symptoms later as adults.
Model 2: Asthma Causes Obesity

Asthma ——> Sedentary——b Obesity
Behavior

Steroid
Use

Obesity is a consequence of asthma and may closely reflect severity of the
disorder. In Model 2, asthma or airway reactivity precedes. obesity. Physical
activity limitation with decreased energy expenditure may tilt the sensitive
metabolic balance towards energy conservation with resultant weight gain.
This may be the most difficult model to refute because of the poor sensitivity
of pulmonary function tests to detect very mild ainNay obstruction, thereby
asthmatics could be classified as non-asthmatic. This error would introduce
differential missclassification as mild asthmatics that could later become
obese classified as non-asthmatic. Airway resistance and specific _
conductance measures on PFTS are particularly good at identifying ainlvay
obstruction and hyperresponsiveness.

The multiple risk factors for obesity further reduce the certainty with which
asthma alone can be determined to lead to excessive weight. In addition,
subnormal lung states can lead to lifestyle changes, e.g., decrease in physical

I.14 posit that obesity

activity behavior and subsequent obesity. Camargo eta
does not occur as a result of changes in physical activity behaviors among

asthmatics. In a multi-variate analysis using a variety of physical activity

45

measures, there was no change in the association between asthma and

obesity at different levels of physical activity.

Model 3: Confounding by a
variable common to both

 

disorders
e.g. ___________.——————-—-> Obesity
Genetic factors Y
Formula (bottle) \,
feeding Asthma

 

 

 

In Model 3, we propose there is a third variable, Y, or a constellation of
variables that predispose to both asthma and obesity, e.g., gene mutation,
depressed hypothalamic pituitary adrenal axis, and adrenergic receptor
responsiveness. This model is largely hindered by existing biological
knowledge on genetic and cellular mechanisms in the pathophysiology of
obesity and asthma. Plausible evidence exists that suggest early childhood
factors such as lack of breastfeeding or beta-adrenergic receptor
polymorphisms may act as confounder of the obesity-asthma link.
Breastfeeding has been linked to lower rates of asthma.70 Breastfeeding,
especially when exclusive, reduces exposure to external antigens during a
period when the immune system is still relatively immature. Additionally,
transfer of immune factors, e.g., lgA, IgG, via breast milk, may yet serve as a
protective mechanism against infections, which can be stimuli for atopy. In a
meta-analysis, Gdalevich et al. identified 12 articles from a MEDLINE search

from 1966-1999 that examined the relationship between breastfeeding and

46

asthma.7o Controlling for age, socioeconomic status, family history of atopy
and parental smoking, they reported a pooled odds ratio of 0.70(0.60-0.81)
between breastfeeding for at least 3 months and subsequent development of
asthma. The protective effect was stronger when the analysis was restricted
to children with a family history of atopy and asthma (OR 0.52).
Breastfeeding is also reported to exert a protective effect on subsequent
“obesity later in childhood.”73 Suggested biological mechanisms include
nutrient composition of breast milk, suckling experience and physiologic
responses to breast milk. For example, breast milk may contain growth
factors that impact adipocyte proliferation. FonnuIa-fed babies are also more
prone to overfeeding, with resultant increase in size of fat cells. Additionally,
breastfed infants compared with formula-fed infants have been shown to
consume much less energy and protein. Von Kries (Germany)72 and Gillman
(USA)73 reported a protective effect using retrospective data. Gillman used a
nationwide cohort of 8186 girls and 7155 boys at ages 9-14 years, who were
offsprings of the participants of the Nurses Health Study. After adjusting for
age, sex, sexual maturity, energy intake, time spent watching television,
physical activity, and socioeconomic status, infants that were mostly or
exclusively breastfed for 3 months were less likely to be obese (OR 0.78;
95%Cl 0.66-0.91) by early adolesCence.
Genetic mutations common to both disorders present another important
plausible mechanism that needs to be further explored. For example, beta

adrenergic influence on ainNay reactivity and fat metabolism has been well

47

documented. Recently there have been studies linking polymorphisms in the
B2 adrenoceptor gene with childhood obesity, asthma in women and physical
activity.74 B2 adrenoceptor Gly16 polymorphism has been associated with
asthma severity while Glu 27 polymorphism has been reported more often in

obese women.”76

Underweight

X > Asthma
Overweight /

 

In Model 4, BMI is depicted as a modifier variable. The effect of Factor X on
asthma is modified by BMI. Some studies have shown that undenlveight,
characterized as BMI less than the 10%tile, is a risk factor for asthma.
Schacter, in a British cohort of children, reported a two-fold increase in
asthma among children below the 10%tile for age and sex.4 (Table 2.1) Lusky
et al. describe Similar findings.11 (Table 2.1) In this model, physical activity
may represent factor X. Undenlveight and obese children have been reported
to exhibit low physical activity patterns, thereby diminishing the amount of full
tidal excursions.”78 Decreased tidal volume sets up the “latch cycle”
described earlier leading to hyperactivity and bronchospasm within the
bronchioles. Model 4 remains poorly developed relative to existing literature
on obesity and asthma, but is important for further study. The modifying effect
of weight on different health conditions suggest that this model is entirely

plausible.

48

Model 5: No Causal
Relationship

 

> Asthma

 

+ Obesity

Model 5 posits there is no relationship between asthma and obesity. The
increased prevalence of asthma among the obese occurs because these are
two conditions that are becoming increasingly more common. Thus, we may
have ascribed to members of a group, characteristics that they do not
possess as individuals. No ecological study on asthma and obesity was
identified in existing literature but there may be a role for such studies. It can
be useful to examine social and cultural variables underlying both disorders;
for example, studing the obesity-asthma link at a period when restricted
actiVity was standard of care for asthmatic patients. Ecological studies can
also provide useful data by focusing on regions that buck the expected trend.
This model has been refuted by multiple studies examining the causal
relationship discussed earlier in Tables 2.1, 2.2 and 2.3.

In addition, trend analysis of BMI and asthma in 28 employment exchange
areas in England and Scotland (National Study of Health and Growth) from
1972-1994 in eight and nine-year old children, obesity did not explain the rise
in asthma prevalence.79 The unadjusted odds ratio per year for trends of

attacks of asthma between 1982-1994 for 9574 boys was 1.09 (95%CI 1.07-

49

1.11) and for 8974 girls 1.09 (95%Cl 1.07-1.12). When adjusted for BMI and
triceps skinfold standard deviation score, the OR remained unchanged.
Although high BMI was associated with asthma, the trend in BMI did not
account for the rise in asthma prevalence. This suggests that obesity may be
a marker for another closely related variable that predisposes to asthma,
which was unmeasured in the study. It will be useful to observe if trend
analysis from different geographical regions concur with this study.
Conclusion: An overview of the conceptual models suggests that the
association between asthma and obesity is complex. Discerning the nature of
the relationship is compounded by methodological difficulties in case and
exposure ascertainment, establishing a temporal association and objective
assessment of potential confounders.

Although levels of biologic evidence provide some coherence to the likelihood
of a relationship, our understanding of the causal pathway underlying the
obesity-asthma link remains speculative, as few studies have addressed
plausible biological mechanisms.

Examining the link between obesity and asthma does not lend itself to
experimental study designs, so it is imperative we strengthen the internal
validity of subsequent observational studies. The choice and framing of the
study question is important, because to adequately establish time order,
incident, not prevalent asthma or obesity, should be used. In addition, using a
combination of clinical, physiologic and questionnaire methods to define

asthma will increase specificity of diagnosis and reduce misclassification.

50

Finally, as many potential causal variables and interactions exist in the
obesity-asthma relationship, further statistical modeling techniques need to be
employed to help determine which variables are likely to be independently
related to the outcome of interest.

The effects of genes, hormones, prenatal and early childhood events,
physical activity, gastroesophageal reflux, and diet in this association are
closely intenNined and present a formidible task to unravel. However, the
rising prevalence and public health impact of both conditions demands further
investigation into the intriguing nature of the obesity-asthma link. Using a
cohort of German children, my thesis will examine the effect of body mass
index on physician diagnosed asthma, bronchial hyperresponsiveness, and
measures of atopy. In addition, I will explore the role of third variables such as

breastfeeding and family history of atopy.

51

CHAPTER 3
DATA AND METHODS

Study Population: The German Child Health and Environment prospective
cohort study is a longitudinal study examining environmental health risks in a
region with waste incinerator and two comparative regions in Hesse,
Germany. There were three consecutive surveys carried out during the winter
and spring in 1994 -1995, 1996 and 1997. Figure 3.1 illustrates the data

collection at each study point.

52

Figure 3.1: Flowchart ofthe various phases ofthe study and the data

used (shaded)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Survey
11’ health
in; s 1994/95 Survey 1996 Survey 1997
N=671
retrospective prospective
4 $
Lung-function " Lung-function I
Test, TW'M” Test, n=sss
N=67l N353 BHR test,
_Weight,lheighkt N=536
Questionnaires Mom Questionnaires
N=67l N=609 N=S67
Blood n =350
IgE and
Eosinophils
Weight &
Birthweight Height at
N=651 Age 4
N=540

 

 

 

 

 

In 1994, parents of second grade children in 18 townships in Hesse, Germany
were invited to participate. Of the 1091 second grade school children ages 7-
10 years attending the schools, 671 (61%) were enrolled in the study. To
lessen selection bias, only schools with at least a 60% rate of participation
were included. The Ethics Committee of General Medical Council, Hamburg,
Germany authorized the study and written consent was obtained.

OUTCOME ASSESSMENT: Physician diagnosis of asthma, history of
wheezing, exercise-induced bronchospasm and asthma symptoms are
selected as health outcomes. Objective measures of atopy (immunoglobulin E

and esinophil count), obstructive indices of pulmonary function (F EV1,

53

FEV1/FVC, FVC, PEF, MMEF) and bronchial hyperresponsiveness to
hypertonic saline were additional outcomes assessed.

Questionnaire: A standardized self-administered questionnaire to collect
data on sociodemographic factors, parental atopy, and history of respiratory
symptoms was completed by the parent. The questionnaire included
questions from the lntemational Study of Asthma and Allergy in Childhood
study (ISAAC). Responses to “Has your child ever had asthma?” and “Has
your child ever had a doctor’s diagnosis of asthma?” were used to define
lifetime prevalence of asthma and physician-diagnosis of asthma respectively.
Atopic eczema and hayfever were based on the question regarding physician-
diagnosis of either disorders. The question related to wheeze was “Has your
child had wheezing or whistling in the chest in the last 12 months?” In
addition, responses to questions regarding history of nocturnal cough,
coughing or wheezing with exercise, and history of physician-diagnosed
wheezy bronchitis were recorded. (Appendix A). Responses to the
questionnaire administered in the first survey 1994/95 were used in the
statistical analysis.

Pulmonary Function Tests: Timing of the last asthma medication and
history of recent respiratory infection was obtained at the onset of testing. All
measurements were performed between 103m and 5:30pm by two
technicians using a Jaeger masterscope 4.1. The instruments were calibrated
daily. Each child performed two forced expiratory maneuvers according to

ATS criteria. The test was considered acceptable if the difference between

54

FVC on both maneuvers was less than or equal to 5%. For purpose of our
analysis FEV1, FEV1/FVC, FVC, PEF and MMEF measures obtained during
the first survey (1994/5) were used. Asthma medications were not
discontinued for the tests.

Bronchial Hyperresponsiveness (BHR) using 4.5% Hypertonic Saline:
BHR was carried out during the third year of the survey in 1997. Subjects
were excluded from the tests if their baseline FEV1 was < than 65% of
predicted values. History of respiratory infection during the last one week was
obtained. BHR was defined as a fall in FEV1 greater than 15%. The bronchial
challenge test was done following standard procedures decribed by Anderson
et al. 8° The 4.5% hypertonic solution was prepared by adding 459m of
dialysis grade sodium chloride to 1000ml of sterile pyrogen free water and
delivered as an aerosolized solution. The mean output of the nebulizer was
2.14mLJmin. During the test, the dose of hypertonic saline was increased
successively by doubling the aerosol inhalation time starting with 30 seconds
up to 8 minutes. One minute after each inhalation, a flow curve was recorded.
The challenge test ended when FEV1 had fallen between 10 - 15%, or after a
cumulative inhalation time of greater than 15.5 minutes and a cumulative
dose of 23ml of hypertonic saline had been used.

lmmunoglobulin E and Eosinophll count: Venous specimen was obtained
from 350 (52%) of the study population during the first year of the study
(1994/95). It was restricted to children whose parents smoked less than 10

cigarettes per day. A Vacutainer was used for blood sampling and about

55

25ml was drawn. Measurements of lgE in serum were made in the Medical,
Alimentary and Veterinary Institute for Research, Middle Hesse, Division of
Human Medicine, Dillenburg, Germany, using radioimmunoassay (CAP,
Pharrnacia, Uppsala, Sweden). The results were provided in kU/L serum.
EXPOSURE ASSESSMENT

Anthropometric Measures: We measured height (m) and weight (kg) for
each child in three consecutive examinations one year apart (approximately
age eight, nine and ten years) and during the lung function tests. In Germany,
parents keep a record book, the Child's Health Card, that includes
documentation of scheduled health check-ups from birth to school age. We
asked the parents to bring their Child's Health Card to a medical examination.
We recorded the child’s height and respective age documented during check-
ups at birth, 4-6 weeks, 3-4 months, 6-7 months, 10-12 months, 21-24
months, and 43-48 months from the child’s health card. The child’s age was
derived from child’s birth date and date when height was ascertained. Body
mass index was calculated by weight(kg) divided by height(m)2. BMI values
were grouped into quintiles for statistical analysis.

OTHER STUDY CHARACTERISTICS

Birthweight was obtained from the child’s medical record. It was divided into
three categories: low birthweight (<25009ms), normal birthweight (2500-
<40009ms) and high birthweight (>40009ms). Duration of total and exclusive
breastfeeding was grouped into 4 week intervals: none, 1-3 wks, 5-8 wks, 9-

12, >12 weeks. When the patient was only breastfed, without introduction of

56

any other form of milk, breastfeeding was defined as exclusive. Passive
smoke exposure was divided into categories based on number of cigarettes
smoked in the household, with the highest category being 20 or more
cigarettes per day. Paternal or maternal history of atopy was positive if either
parent reported a history of atopic dermatitis or hayfever. -

Statistical analysis: To address the study aims, the study population was
analyzed in three subgroups based on the presence of missing data (Figure
3.2). Of the 671 children enrolled in the study, 515 subjects had detailed
information on anthropometric measures and potential confounders. The total
population used for analysis was 515, 55% of the original cohort eligible for
the study. The subgroups were: 1) total population was restricted to 515
children; 2) 480 children with PFT results and bronchial hyperresponsiveness;

and 3) 269 children with eosinophil counts and immunoglobulin E levels.

57

Figure 3.2 Defining study population and subgroups used for analyses

Total population

 

[ =671 J

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Subgroups ,
Goup 1 variables (Group 1 W [Group I \
variables variables
BMI @ 4
BMI @ 7 flag
, Birthweight total PFT m
Definbryg Breastfeeding age PFT
vana es Cigarette smoke IgE
for subgroups E0 3 i no phil
L J C... 9
Study population .

   

Indicator, or dummy, variables were created for age, sex, BMI at ages 4 and 7
years, birthweight, exposure to passive smoking and breastfeeding. Group
comparisons were carried out by cross tabulations and Chi (x2) test of
proportions. To examine predictors of asthma and bronchial
hyperresponsiveness, logistic regression with BMI at age 4 and 7, sex,
breastfeeding and family history of atopy in the model was carried out. To be
more inclusive, children with either a wheezy bronchitis, physician-diagnosed
asthma, exercise-induced bronchospasm or positive hyperresponsiveness
were grouped together and logistic regression analysis also performed.

Because the distributions of lgE and eosinophil count were not normal, the

58

geometric mean, median, and 5th and 95th percentiles were used. Those
variables that were associated with both the outcome and exposure variables,
either in our data or from other literature, were re-entered into a regression
analysis along with interactions between the exposures. The variables that
remained in the final model were BMI at ages 4 8r 7, age, birthweight,
breastfeeding, sex, height, maternal and paternal atopy, history of respiratory
tract infection a week prior to PFT and passive smoke exposure. Categories
of variables have been described earlier in the chapter and are presented in
Table 4.1.These variables were also analyzed as potential indicators of
eosinophil count, immunoglobulin E and pulmonary function measures in a
linear regression model. All data analyses were carried out using the SAS

statistical package (SAS institute, Cary, NC) version 8.

59

CHAPTER 4
RESULTS

Study population characteristics: The median age was 8.2 years during
the first survey (1994/95) and males constituted about 55% of the population
in all subgroups. In the total population, the prevalence of physician-
diagnosed asthma, hayfever and atopic eczema was 5.3%, 8% and 26%
respectively. Eighty-seven percent of the children were born with normal
birthweight (2500-40009m), 4.7% with low birthweight (<25009m) and 8.7%
with high birthweight(>40009m). Thirty-six percent had been breastfed for at
least 12 weeks, whereas 20.2% were not breastfed at all. About half of the
children had exposure to environmental tobacco at home. One-quarter and
23% of the fathers and mothers, respectively, reported a history of atopy
(Table 4.1).
The subgroup with data on IgE and eosinophil count differed slightly from the
total population (Table 4.1), with a lower number of children reporting tobacco
exposure (32%) and 12% with birth weights above 4000 grams versus 8.7%.
Children whose families smoked >= 10 cigarettes/day had been excluded
from obtaining IgE and eosinophil counts. They were also more likely to have
been breastfed for longer than 12 weeks (43% vs. 36%) than the total
population. However, none of the differences were statistically significant.
Relationship of BMI quintiles to US growth chart: Mean BMI at age 4 was
16.8 (range 126-242) and 16.2 (12.8-30.4) at age 7. Children with BMI

greater than 95th percentile for age and sex are defined as obese. The lowest

60

value of BMI in the 5th quintile of BMI was comparable to the 95th percentile

cut-off for BMI on the 2000 US growth chart. For age 7, the minimum value in

the 5th quintile was 20.5 compared to 20.08 for boys and 20.78 for girls at the

95% in the US growth chart. Likewise, the minimum 5“1 quintile value at age 4

for boys and girls was similar to the US growth chart at 95th percentile (17.9

vs. 17.2) for boys and (17.9 vs. 17.4) for girls.

Table 4.1 Characteristics of study population

 

 

 

 

 

 

 

 

 

 

 

Characteristic Total PFT lgE/Eosinophll
n=515(%) N=480(%) N=259(%)
Median Age 8.2 8.2 8.2
Male sex 283 (55) 259 (54) 148 (57)
Birthweight (gms)
<2500 24 (4.7) 19 (4.0) 7 (2.7)
z2500<4000 446 (86.6) 416 (86.6) 221 (85.0)
3 4000 45 (8.7) 45 (9.4) 31 (12.0)
Wheezy bronchitis 48 (9.4) 43 (9.0) 24 (9.3)
Physician-Diagnosed Asthma 27 (5.3) 25 (5.2) 15 (5.8)
No Physician-Diagnosed 434 (84.4) 409 (85.7) 218 (84.8)
Asthma ,
BMI at age 4 (quintiles)
<15.61 102 (19.8) 96 (20.0) 50 (19.3)
15.64-<16.44 104 (20.2) 93 (19.4) 53 (20.5)
16.44-<17.19 96 (18.6) 103 (21.5) 57 (22.0)
17.19—<17.99 107 (20.8) 88 (18.3) 55 (21.2)
217.99 106 (20.6) 100 (20.8) 44 (7.0)
BMI at age 7
(quintiles)
>15.98 105 (20.4) 95 (19.8) 58 (22.4)
15.98-<17.05 100 (19.4) 95 (19.8) 44 (17.0)
17.05—<18.57 107 (20.8) 89 (18.5) 47 (18.1)
18.57-<20.54 96 (18.6) 99 (20.6) 54 (20.9)
2 20.54 107 (20.8) 102 (21.2) 56 (21.6)

 

61

 

Table 4.1 (cont’d.)

 

 

 

 

 

Total breastfeeding (wks)

None 104 (20.2) 89 (18.5) 35 (13.5)
0 <4 98 (19.0) 93 (19.4) 41 (15.8)
_>_ 4>8 68 (13.2) 66 (13.7) 36 (13.9)
3 8<12 58 (11.3) 57 (11.9) 35 (13.5)
212 187 (36.3) 175 (36.5) 112 (43.2)
Environmental smoke

exposure 282 (54.8) 260 (54.2) 86 (32.1)
Yes 233 (45.2) 220 (45.8) 176 (67.9)
No

Maternal history of atopy

Yes 120 (23.3) 107 (22.3) 59 (22.8)
No 395 (76.7) 373 (77.7) 200 (77.2)
Paternal history of atopy

Yes 106 (20.6) 98 (20.4) 64 (24.7)
No 409 (79.4) 382 (79.6) 195 (75.3)

 

 

 

 

 

Birthweight was correlated with BMI at age 4 years r2 =0.14 (p0.0003) and
BMI at age 7 years r2 =0.18 (p< 0.0001) while body mass index at age 4
years Was moderately correlated with BMI at age 7 years r2 =0.22 (p
<0.0001). Total breastfeeding was not correlated with BMI at age 4 r2=0.005
(p=0.90) or 7 years r2=0.005 (p=0.90), but was correlated with birthweight r2:
0.14 (p=0.0026). (Appendix B).

The relationship between BMI and 1) asthma and asthma symptoms; 2)
pulmonary function test; and 3) Eosinophll count, IgE and bronchial
hyperresponsiveness (BHR) will now be examined.

Asthma symptoms and diagnosis: The prevalence of physician-diagnosed
asthma was 5.3% and boys were more likely to be asthmatic than girls OR
3.1 (95%CI 1.2-8.2). Breastfeeding was protective for physician-diagnosed

asthma (OR 0.7 95%Cl 0.5-0.9) .

62

Children with physician-diagnosed asthma were more likely to have hayfever,

27% vs. 3.5% (p=0.0001), atopic eczema, 8.7% vs. 4% (p=0.03) and

wheezing without infection, 58% vs. 3.4% (p=0.0001). Exposure to tobacco

smoke at home was not a risk factor for asthma (p=0.17).

Table 4.2: Relationship of BMI, breastfeeding and family history of atopy

 

 

 

 

 

 

to asthma
Variables Physician- History of asthma, or Any Asthma
Diagnosed exercise-induced symptom
N = 515 Asthma bronchospasm, or N= 141
N=27 wheezy bronchitis or
BHR =84

OR 95% Cl OR 95% Cl OR 95% Cl
BMI at age 4 0.9 0.3-3.8 0.8 0.4-1 .4 0.5 0.3-1.0
lst vs 5th quintile
BMI at age7 1.7 0.4-5.9 0.7 0.4-1.4 1.8 0.9-3.5
lst vs 5th quintile
Total 0.3 0.1-0.8 1.5 0.8-2.7 0.7 0.4-1.8
breastfeeding ‘
lst vs 5th quintile
Maternal atopy 1.5 0.6-3.7 1.1 0.7-1.9 1.4 0.9-2.3
lst vs 5th quintile .
Paternal atopy 2.3 1.0-5.4 0.6 0.3-0.9 1.2 0.7-1.8
lst vs 5th quintile

 

 

 

 

 

The odds of physician-diagnosed asthma did not increase with BMI quintiles

at age four or seven (Table 4.2).

Stratifying by sex, the result was unchanged. A positive effect of only female

sex on the association between asthma and obesity have been reported in

certain studies. However, testing for an interaction between sex and BMI on

asthma in our study population, there was no significant relationship (p=0.92).

Using a logistic regression model, controlling for sex, age, maternal and

63

 

paternal atopy, passive smoke exposure and breastfeeding, neither BMI at
age 4 or 7 was a risk factor when asthma was redefined with the more
rigorous classification described in Chapter 3.

When children with either wheezy bronchitis, physician-diagnosed asthma,
exercise-induced bronchospasm or positive hyperresponsiveness were
grouped together and used as the dependent variable in the regression
model, there was no association with BMI at age four or seven years. (Table
4.2).

Reanalyzing the data at age 7 years using US growth chart BMI percentiles
as a reference, the result remained unchanged (Table 4.3).

Table 4.3: Association between asthma and BMI using 2000 US growth
charts as reference.

 

 

 

BMI Percentile1 Physician-Diagnosed Asthma
OR 95% CI.

Males

_>_ 95 0.7 (009-58)

2 85 0.56 (0.16-1.96)

5 10 0.7 (0.27-1.69)

_<_ 5 1.5 (043-55)
Females

Z 95 2.7 (0.11-46.4)

_>_85 1.1 (0.1-11.26)

_<_10 2.9 (0.4-21.1)

_<_5 1.1 (0.1-10.9)

 

 

 

 

 

1Total population >95%tile N= 49;(9.5%) >85-95%tile N: 169 (32%)
Asthma symptoms: Twenty-seven percent of the total population reported at
least one asthma symptom. Of the total population of 515 children, five

percent had wheezing with exercise, while three percent and 16% reported

64

wheezing and dry cough without upper respiratory tract infection respectively.

A larger proportion of children in the 4th and 5th quintile had a diagnosis of

wheezy bronchitis (Table 4.4)

Table 4.4: Percentages of atopic conditions and asthma symptoms

within BMI quintiles for age 7 years

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Symptoms 1St 2“d 3'“r 4'“ 5'"
Quintile Quintile Quintile Quintile Quintile
N=105 N=100 N=107 N=96 N=107
_Ale 7 years

Asthma 4.7 6.0 3.8 5.2 6.6
Wheezy 8.4 6.9 7.6 9.9 12.3
bronchitis
Hay Fever 8.5 4.0 8.4 8.3 6.5
Atopic 25.7 25.0 33.6 29.1 20.5
Eczema
Dry Cough 11.4 24.0 17.7 13.5 14.9
Cough after 10.4 16.1 16.8 15.7 19.8
exercise
EIB* 3.8 6.0 5.6 5.2 9.4
History of 12.3 10.0 10.3 6.2 12.2
wheeze
Nocturnal 23.8 23.0 32.7 22.1 28.9
cough
At least one 20.9 30.0 3l.7 23.9 29.9
asthma
symptom
Number of
asthma
symptoms" 79.0 70.0 68.2 76.0 70.0

0 9.5 16.0 25.2 15.6 15.8

1 6.6 6.0 3.7 4.1 7.4

2 3.8 5.0 1.8 3.1 6.5

3 0.9 3.0 0.9 1.0 0.0

4

 

* Exercise induced bronchospasm
** Number of positive responses to any of the asthma symptoms (See Appendix A)

65

 

Children with exercise-induced wheezing were more likely to be in the 5th
quintile at age 7 years, 8.5% vs. 3.5% (p=0.05). However, the prevalence of
hayfever, nocturnal cough and history of wheezing did not rise significantly
with BMI at age 4 or 7 (Figure 4.1).; At age 7 years, children with EIB had
more symptoms of nocturnal cough, 50% vs 3.5% (p=0.0001) and a history of
hayfever, 12.5% vs 4.5(p=0.005) as illustrated in Table 4.4. They were also

more likely to have a positive BHR 15.8% vs. 3.5% (p=0.0001).

 

Flgure 4.1: Prevalence of Asthma, Hayfever, Exercise
Induced Bronchospasm [EIB] and Asthma syrrptoms by
Body mass Index at 7 years
*p<005

 

_a _a _L _a
N h C)
l L I

Prevalence (%)

 

 

L l

 

 

 

ore-501060

 

2 3 4
Quintiles of BMI

 

 

 

There was no significant difference in prevalence of asthma symptoms and

BMI at age 4 and 7 (Appendix C lists the percentages of asthma symptoms

within BMI quintiles at age 4 years).

66'

Pulmonary function tests: Spirometric tests for obstructive lung disease
such as FEV1, FVC, FEV1/FVC, FEF50 and PEF were selected for analysis.
Abnormalities in these measures, as discussed in Chapter 1, typically reflect
ainrvay obstruction, which is the underlying pathophysiologic process for
asthma. To control for any confounding effects, multiple logistic regression
analysis was used to explore the variation of pulmonary function measures by
BMI, while controlling for age, sex, maternal and paternal atopy, passive
smoke exposure, breastfeeding and history of recent respiratory infection
(Table 4.5). PEF, F EV1, FVC and FEV1/FVC values were significantly higher
in males than females. BMI at age 4 3 17.9 (5th quintile) was associated with
a decrease in FEV1/FVC (p=0.01) and PEF (p=0.01) when compared with the
group with lowest BMI. At age 7, there was an increase in FVC for BMI >
20.5, (5th quintile p=0.04) but no other Significant differences in FEV1,
FEV1/FVC or PEF were found.

There was no consistent trend in the protective effect of breastfeeding on
FEV1, FEV1/FVC, MMEF or PEF. Compared with no breastfeeding, total
breastfeeding for a duration of 8 weeks was associated with a significant
increase in PEF (p=0.02) and FEV1 (p=0.03), while breastfeeding for greater
than 12 weeks was associated with an increase in FEF50 (p=0.02) and
FEV1/FVC (p=0.0004). The results were similar when the analysis was

carried out for exclusive breastfeeding.

67

 

 

 

 

 

 

0:.0 00.0 00.0 :00 00.0 E: 00.0 :00 0:.0 00.0 00.0%
0:.0 :00 0:0 00.: 00.0 00.: 00.0 00.0 00.0 30 009.00:
00.0 00.0 00.0 00.: 00.0 00.: 00.0 00.0 0:.0 00.0 00:00.:
0:.0 :00 :00 00.: 00.0 00.: :00 0:.0 00.0 00.0 00.::v-00.0:
I 00.0 I. 00.: .l 00.: .l :00 I 00.0 00.0:A
.000"... .02.“: 8:08 002.0: :00"... .02.“: f0"... .03.": .0002. 00...... 00.00.00.
. 20.5.0
:00 00.0 :00 000 00.: :00 00.0 :00 00.0 00.::N
:0 0.0 00.0 00.: 00.0 :0: 00.0 0:.0 0:0 00.0 00.::v-0:.:
:00 . 00.0 0:0 00.: 00.0 ::.: 00.0 :00 :0 00.0 0:.::v-$..0:
00.0 00.0 00.0 00.: :00 00.: 00 00.0 00.0 00.0 30:40.0:
I 00.0 I 00.: l... 00.: I. 00.0 I :00 :0.0:v
.000ua .02.“: 5.0%. 003.0. 03.1.0 003.0. .0008 .02.": .0008 .02.": 30.00.00.
30.20
0000.0 00.0 00.0 00.: 00.0 :0: 00.0 :00 :0 3.0 mm
:0 0.0 :0 00.: :00 0A.: 0.0 :00 0:.0 00.0 0V0N
:0 00 00.0 00.0 00.0 00.: :0 0:.0 00.0 0.0 0veN
000.0 00.0 00.0 00.: 00.0 E: 00.0 00.0 00.0 00.0 ev-0
.I 00.0 I. 00.: l. 00.: i. 00.0 .I 00.0 0:02
80001030. .0000 .02.“: 60.0»0 002.0. 50".. .02.“: 00.0% .02.“: $0.20
8600:0005
00.0 00.: 00.: 0:.0 00.0 0.08“.
:00 :00 00.: 0:.0 00.0 0.0.2
.0008 .09.“: :000ua .02.“: $808 .09.": .308 002.0. .000"... .02.": .80
009.0 009.0 000:-.. 009.0 0000.0
m:_m>rm cam—twi— 0:_m>rn_ cmmerd m:_m>rn_ cmmEmJ 0:_m>r& :mmEmA ¢:_m>rn_ cmmEmr— omwl Z
o>n_.:>m.n. ”.00 o_aa.a>

 

 

 

 

 

 

 

om “Em

20:02:“. Deco—=3.“ :o 200»
n new .0 00mm an .55 can 553030005 :0 «00.:—m 05 no: 02:03. 5.00053. .0023 0.02:! .06 030...

68

Atopy and Bronchial hyperresponsiveness: The strong association between
asthma and atopy, sedentary behavior with increased exposure to indoor
allergens as well as the inflammatory mechanisms of adipose tissue, suggests
atopy may be a plausible biological mechanism to explain the asthma obesity
link. Table 4.6 summarizes the effect of BMI at age four years, BMI at seven
years and breastfeeding on bronchial hyperresponsiveness. A total of fifteen
percent of the study population tested for bronchial hyperresponsiveness using
4.5% hypertonic saline had a positive result.

Table 4.6: Effect of BMI on Bronchial Hyperresponsiveness

 

 

 

 

 

Variable Bronchial
hyperresponsiveness
N= 536
Odds
Ratio 95%C.l
BMI quintiles (wt/htz) Age 4 years
<15.61 1.0 (ref)
15.64-<16.44 0.9 (0.4-2.4)
16.44-<17.19 0.8 (0.3-2.2)
17.19-<17.99 1.4 (0.5-3.4)
21799 1.2 (0.5-3.0)
BMI quintiles (wt/htz) Age 7 years
>15.98 1.0 (ref)
15.98-<17.05 1.2 (0.5-3.1)
17.05-<18.57 0.6 (0.2-1.4)
18.57-<20.54 0.7 (0.3-1.7)
2 20.54 0.8 (0.3-3.2)
Total breastfeeding (wks)
None 1.0 (ref)
0-<4 0.9 (04-22)
24<8 0.6 (02-19)
23412 1.0 (0.4-2.6)
212 0.4 (0.2-0.9)

 

 

69

 

Children who had a physician-diagnosed asthma (OR = 1.2, 95%CI 1.07- 1.32),
hayfever (OR = 1.3, 95%Cl 1.0- 5.2), exercise-induced bronchospasm
(p=0.0001) and wheezing without evidence of respiratory infection (p =0.001) but
not lgE (p=0.06), were more likely to have a 15% fall in FEV1. Breastfeeding for
greater than 12 weeks had a protective effect against bronchial
hyperresponsiveness (OR = 0.4, 95% CI 0.2-0.9). There was no association
between bronchial hyperactivity and BMI at age 4 (p=0.54) and BMI age 7
(p=0.68).

Children with asthma were more likely to have elevated IgE levels (p=0.0004).
Table 4.6 presents geometric means of eosinophil count and lgE stratified by
BMI at ages 4 and 7 and breastfeeding. Log transformation of lgE and eosinophil
was used for analysis. There was no significant association of BMI with serum

eosinophil and lgE.

7O

Table 4.7: Effect of BMI on lmmunoglobulin E and Eosinophil Count.

 

 

 

 

Variable lgE (KU/L) N=259 Eosinophil (absolute count/nl)
N=259
Geometric P-value Geometric P-value
mean (t-test) mean (t-test)
BMI uintiles
(wt/h ) Age 4 (F-test p=0.30) (F-test p=0.34)
years
<15.61 80.58 --- 251.87 ---
15.64-<16.44 80.87 0.05 225.83 0.35
16.44-<17.19 57.95 0.72 281.90 0.55
17.19-<17.99 61.10 0.85 228.86 0.94
217.99 49.25 0.51 258.12 0.08
BMI uintiles
(wt/h ) Age 7 (F-test p=0.92) (F-test p=0.21)
years
>15.98 65.44 ---- 244.67 ---
15.98-<17.05 47.84 0.53 181.98 0.79
17.05-<18.57 64.03 0.41 267.83 0.21
18.57-<20.54 83.15 0.41 286.66 0.03
_>. 20.54 68.17 0.43 277.07 0.60
Total
breastfeeding (F-test p=0.09) (F-test p=0.‘37)
(wks)
None 75.37 ---- 273.51 ----
0-<4 35.78 0.10 205.46 0.95
24<8 64.26 0.01 240.53 0.05
23412 70.88 0.18 254.02 0.39
212 92.52 0.62 - 275.87 0.72

 

 

 

 

 

Only breastfeeding for less than 8 weeks was significantly associated with lower

lgE levels (p=0.01) and eosinophil count (p=0.05).

71

 

CHAPTER 5

DISCUSSION

A high BMI was not associated with an increasing prevalence of asthma in this
population of 515 German children aged 7-8 years controlling for potential
confounders such as passive smoke exposure, sex, breastfeeding, birthweight,
maternal and paternal atopy. There was no effect when stratifying by sex and no
interaction between sex and BMI. The lack of association persisted for BMI at
age 4 and subsequent history of asthma at ages 7-8 years. To be more inclusive,
children who had a physician-diagnosis of asthma or exercise-induced asthma or
wheezy bronchitis or any asthma symptom were grouped together. Logistic
regression using this outcome measure still did not Show a relationship between
BMI both at ages four or seven years.

The longitudinal nature of the study design allows us to ascertain BMI at age
four, prior to determination of a history asthma, bronchial hyperresponsiveness,
obstructive pulmonary indices and immunoglobulin level. Regardless of whether
the children were in the heaviest quintile at age 4 or 7, the results were
unchanged. Thus, there was no relationship between obesity and asthma.

Our findings are in agreement with results reported by Brenner21 and Lusky et
al.11 and others who found no association between asthma and obesity but
discordant with some investigators?“ 12402223 In a British study carried out
between 1982-1994, Chinn et al. found that an increase in body mass index

could not explain the rise in asthma.79 The odds ratio per year for asthma across

72

 

the time period of the study remained unchanged when adjusted for BMI. Their
findings suggest the obesity-asthma link may not be causal but rather reflect the
effect of recent lifestyle changes. It will be important to re-examine the causal
link in different populations and geographic location with varying patterns of
lifestyle behaviors and physician diagnostic patterns, as most of the
observational studies reporting a postive association have been carried out either I
in the US, UK or Canada.

Most studies that found apositive association between asthma and obesity used

a history of asthma either obtained by self-report or through medical chart review

 

and medication use to define their population. In our analysis, self-reported
history of asthma symptoms, asthma diagnosis and objective measure of BHR
were all used as outcome measures, thus reducing the chance of
misclassification. But there was still no association found with various
permutations of diagnostic definitions. It is improbable that physician diagnostic
patterns may account for the lack of association found Since we were unable to
report a link between objective measures of ainrvay reactivity and BMI.

An association between asthma and obesity may also have been attenuated by
the high rate of breastfeeding in the sample. About 80% had been breastfed, with
36% breastfed for longer than 12 weeks. Breastfeeding has been reported to
have a protective effect on both asthma and obesity. In a meta-analysis on
breastfeeding and asthma by Gladevich et al. they reported a summary OR of
0.70 (95%Cl 0.60-0.81).7o Indeed, we found a protective effect of breastfeeding

on physician-diagnosed asthma in our study. However, when breastfeeding was

73

controlled for in the model, the lack of association between asthma and BMI
persisted.

In contrast, exercise-induced bronchospasm (EIB) was related to BMI at age 7
years. EIB could account for only part of the reported association between
asthma and obesity. Kaplan et al. reported increased frequency and severity of
EIB among 6-10 year old obese non-asthmatic children compared with their non-
obese counterparts.450 The mechanism whereby EIB occurs is thought to occur
from consequences of heating and humidifying large volumes of air during
exercise. It is unknown whether this process is different in ovenlveight persons.
An alternate explanation may be impaired chest wall mechanics reported in
obese individuals with a resultant decrease in end tidal volume leading to slow
cycling rates in the actin-myosin complex. The subsequent bronchial
hyperreactivity could then be exacerbated by exercise.

A high BMI was not associated with a positive BHR test in our study population.
Similar findings were reported in a study of 2,842 children aged 13-14 years who
underwent exercise-challenged testing for bronchial hyperresponsiveness. The
investigators found a 15% fall in FEV1 was significantly associated with lower
age, female sex, high socioeconomic level and attending a private school, but not
obesity.81 There is a discrepancy between a positive effect of BMI on exercise-
induced bronchospasm yet not on bronchial hyperactivity since both conditions
are closely related. Failure to discontinue asthma medications prior to testing

may account for this finding.

74

The spirometric pulmonary function measures PEF, FEF50 , FEV1, FEV1/FVC
and FVC volumes were all higher in males as has been reported in the literature.
High BMI measured at age four years resulted in dimunition in pulmonary
function measures at age 7-9 years. At age 4, after controlling for confounders,
children in the highest quintile had lower values for PEF and F EV1/FVC
measured but not F EV1. FEV1/FVC and FEF50 are the most important measure
in distinguishing an obstructive‘impain’nent while PEF is best used for monitoring.
Because BMI at age 4 and 7 were correlated, this finding may reflect persistence
of ovenlveight in the same individuals. A 3% decline in FEV1/FVC and 0.23L
reduction in FEF50 indicates obesity has some effect on airway obstruction. This
effect most likely would have been more pronounced had asthma medications
been discontinued for the PFTS. The increase of 0.1 L in FVC among children in
the 5th quintile is contradictory, but this may reflect the weakness of body mass
index in distinguishing between lean body mass and fat. Children Who are more
physically active are more muscular and have larger vital capacity.

Only pulmonary function test (PFT) results from the first survey were used in the
analysis. Serial PFT measures would have been more informative. Airflow
obstruction in asthma is reversible, thus a Significant number of asthmatics can
have normal PFT tests at a single point in time which may not truly reflect the
extent of obstruction. In addition, the tests were done between 10 am-5:30 pm, a
period of time during the day that airflow obstruction is likely to be at its least.
Atopy is strongly associated with asthma. About 80% of asthma in children and

adults has been linked to atopic sensitization in clinical studies. It has been

75

hypothesized that obese subjects spend more time indoors, thus increasing their
susceptibility to indoor allergens and sensitization. This hypothesis along with the
strong association between asthma and atopy and inflammatory mechanisms of
adipose tissue, suggest atopy may be a plausible biological mechanism to
explain the asthma obesity link. However BMI was not associated with the
measures of atopic sensitization, hayfever, atopic eczema, total immunoglobulin
E and eosinophil count assessed in our analysis. The findings are supported by
von Mutius et al.7 that the association between BMI and asthma is not based on
an atopic mechanism. Using the NHANES III data, they found BMI was not
related to eosinophil count or Skin prick tests to common allergens including
dustmite, gennan cockroach, cat, grass, ragweed, Altemaria and peanuts.

The German Child Health and Environment study in Hesse, Germany contains
detailed information on asthma and risk factors in a large cohort. Availability of
various measures for asthma provide good data to minimize misclassification and
pulmonary function tests and BHR tests were carried out according to the ATS
protocol by experienced technicians.

Another strength is its longitudinal nature, which allows for assessment of
causality. Shortcomings exist in this dataset and some have been mentioned
earlier. Only 61% of total population eligible participated in the study because of
available data and refusal to participate (total population of 515 from the 1091
eligible children). However, there were no majordifferences between the groups
that would inherently affect the results in either direction. A major confounder

like breastfeeding was obtained by self-report more than 5 years following the

76

event; but it is unlikely that mothers of obese children would report any
differently. Finally, based on a prevalence of 5.7% of asthma in our study
population of 515 children, we may not have had enough power to detect a
significant association between asthma and BMI.
In conclusion, in a longitudinal study of German school children, BMI is not an
independent risk factor for asthma, but is positively associated with exercise-
induced bronchospasm. In addition, BMI is not related to lgE, eosinophil count or
bronchial hyperresponsiveness. The decline in ainlvay obstructive parameters at
age 7 years with increasing BMI measured at four years but not at 7 years
provides important information that obesity plays a role in airway hyppereactivity
that may not be mediated through impaired chest wall mechanisms.
Understanding the pathophysiologic mechanism underlying the effect of BMI on
exercise-induced bronchospasm is important as this condition is becoming
increasingly common. Further studies, with emphasis on specificity of exposure
and outcome ascertainment, are needed to elucidate the relationship between

asthma and obesity.

77

APPENDIX A

Questions on asthma and symptoms

 

 

9’ 9‘ :PP’N.‘

Has your child ever had a doctor’s diagnosis of asthma?

Has your child ever had a doctor’s diagnosis of atopic dermatitis?
Has your child ever had a doctor’s diagnosis of hayfever?

Has your child had wheezing or whistling in the chest in the last 12
months?

Has your child had a dry cough, apart from a cough associated with a
cold or chest infection? '

Has your child had attacks of shortness of breath or dyspnea in the
last 12 months?

Has your child had cough or Shortness of breath at night or early in the
morning in the last 12 months? ‘

Does your child have wheezing during or after exercise in the last 12
months?

Does your child have coughing during or after exercise in the last 12
months?

 

78

 

Spearman’s correlation coefficients for selected variables.

APPENDIX B

 

 

 

 

 

 

 

 

 

Variable Birthweight BM|@ 4 yrs. BMI @ 7 yrs. Total
rank rank rank breastfeeding
coefficient coefficient coefficient rank
p-value p-value p-value coefficient
p-value
Birthweight 1.00 0.19 0.14 0.14
<.0001 0.0003 0.0003
N =554 N =632 N =617
BMI @ 4 yrs. 1.00 0.22 0.005
<. 0001 0.90
N=547 N=531
BMI @ 7 yrs. 1.00 0.005
<0.9002
N=619

 

79

 

APPENDIX C

Percentages of atopic conditions and asthma symptoms within BMI
quintiles for ages 4 years

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

sym ptoms 1 St 2nd 3rd 4th 5th
Quintile Quintile Quintile Quintile Quintile
N=102 =104 N=96 N=107 N=106

Age 4 years
Asthma 4.9 5.8 6.2 4.7 4.7
Wheezy 5.3 6.2 7.6 16.5 13.6
bronchitis
Hay Fever 7.8 4.8 2.0 11.3 9.4
Atopic 29.4 28.8 25.0 25.2 25.4
Eczema
Dry Cough 14.7 13.4 19.7 18.6 15.0
Cough after 19.6 18.6 13.6 17.7 9.4
exercise
EIB 4.9 8.6 5.2 8.4 2.8
History of 6.8 12.5 10.5 11.2 10.4
wheeze
Nocturnal 30.3 25.0 27.3 27.1 21.7
cough
At least one 33.3 26.9 25.0 28.9 22.6
asthma
symptom
# of asthma
symptoms

0 66.6 73.0 75.0 71.0 77.3

1 24.5 13.4 15.6 16.8 12.2

2 2.9 5.7 8.3 5.6 5.6

3 4.9 3.8 1.0 6.5 3.7

4 0.9 3.8 0.0 0.0 0.9

 

8O

 

10

11

 

 

BIBLIOGRAPHY

. Asthma Prevention Program, Center for Disease Control; National Center for
Environmental Health 2000.

. MMWR, April 24, 1998/47 (SS-1); 1-28. Surveillance for Asthma, United
States, 1960-95.

. Chen Y, Horne SL, Dosman JA. Body weight and weight gain related to
pulmonary function decline in adults; a six-year follow up study. Thorax.
1993;48:375-80.

. Chinn S, and Rona RJ. Can the increase in body mass index explain the
rising trend in asthma in children? Thorax. 2000;56(11):845-50.

. Schwartz J, Gold D, Dockery DW, Weiss ST, Speizer FE. Predictors of
asthma and persistent wheeze in a national sample of children in the United
States. Am Rev Respir Dis. 1990;142:555-62.

. Figueroa-Munoz Jl, Chinn S, et al. Association between obesity and asthma
in 4-11 year old children in the UK. Thorax. 2002;56(2):133-7.

. von Mutius E, Schwartz J, Neas LM, Dockery D, Weiss ST. Relation of body
mass index to asthma and atopy in children: the National Health and Nutrition
Examination Study Ill. 2001 ;56(1 1 ):835-8.

. Castro-Rodriguez JA, Holberg CJ, Morgan WJ, Wright AL, Martinez FD.
Increased incidence of asthma like symptoms in girls who became overweight
or obese during the school years. Am J Respir Crit Care Med.
2001;163(6):1344-9.

. Celedon JC, Palmer LJ, Litonjua AA, Weiss ST, Wang B, Fang Z, Xu X. Body
mass index and asthma in adults in families of subjects with asthma in
Anqing, China. Am J Respir Crit Care Med. 2002;164(10 Pt 1):1835-40.

10.von Kries R, Hermann M, Grunert VP, von Mutius E. Is obesity a risk factor in

childhood asthma? Asthma. 2002;56(4):318-22.

11.Lusky A, Barell V, Lubin F, Kaplan G, Layani V, Shohat Z, Lev B, Wiener M.

Int J Epidemiol. 1996;Aug;25(4):829-34.

81

12.Jang AS, Son M, Chooi IS, Koh Yl. High body mass index is associated with
wheezing among older adults living in high-altitude area in Korea. J Korean
Med Sci. 2002;Aug;17(4):479-82.

13. Sin DD, Jones RL, Man SF. Obesity is a risk factor for dyspnea but not for
airflow obstruction. Arch Intern Med. 2002;Jul 8:162(13):1477-81.

14. Camargo CA Jr, Weiss ST, Zhang S, Willett WC, Speizer FE. Prospective
study of body mass index, weight change, and risk of adult-onset asthma in
women. Arch Intern Med. 1999;159:2582-2588.

15. Shaheen SO, Sterne JA, Montgomery SM, Azirna H. Birth weight, body mass
index and asthma in young adults. Thorax. 1999 May;54(5):396-402

16.Chen Y, Dales R, Tang M, Krewski D. Obesity may increase the incidence of
asthma in women, but not in men: longitudinal observations from the
Canadian National Population Health Surveys. Am J Epidemiol.
2002;155(3.):191-7.

17.Xu B, Pekkanen J, Laitinen J, Jarvelin MR. Body build from birth to adulthood
and risk of asthma. Eur J Public Health. 2002;Sep;12(3):166-70

18.Gennuso J, Epstein LH, Paluch RA, Cemy F. The relationship between
asthma and obesity in urban minority children and adolescents. Arch Pediatn'
Adolesc Med. 1998;Dec;152(12):1197-2000.

19. Luder E. Melnik TA, DeMaio M. Association of being ovenlveight with greater
asthma symptoms in inner city black and Hispanic children. J Pediatr.
1 998;Apr;1 32(4):699-703.

20.Young SY, Gunzenhauser JD, Malone KE, McTieman A. Body mass index
and asthma in the military population of the northwestern United States. Arch
Intern Med. 2002;161(13):1605-11.

21 . Brenner JS, Kelly CS, Wenger AD, Brich SM, Morrow AL. Asthma and obesity
in adolescents: is there an association? J Asthma. 2001;38(6):509-15.

22. Belamarich PF, Luder E, Kattan M, et al. Do obese inner city children with
asthma have more symptoms than non-obese children with asthma?
Pediatrics. 2000;Dec;106(6):1436-41 .

23.Guerra S, Sherrill DL, Bobadilla A, Martinez FD, Barbee RA. The relation of

body mass index to asthma, chronic bronchitis, and emphysema. Chest.
2002;Oct;122(4):1256-63.

82

24. lnselman LS, Milanese A, Deurloo A. Effect of obesity on pulmonary function
in children. Pediatr Pulmono. 1993;16:130-7.

25.Visser M, Bouter LM, McQuilian GM, Wener MH, Harris TB. Elevated c-
reactive protein levels in ovenlveight and obese adults. JAMA. 1999;Dec
282(22):2131-5.

26.Gonen B, O’Donnell P, Post TJ, Quinn TJ, Schulman ES. Very low-density
lipoproteins trigger the release of histamine from human basophils.
Biochemical et Biophysical. 1987;917:418-24.

27. Black PN, Sharpe S. Dietary fat and asthma: is there a connection? Eur
Respir J. 1997;Jan;10(1):6-12.

28.Loffreda S, Yang SQ, Lin HZ, Karp CL, Brengman ML, Wang DJ, Klein AS,
Bulkley GB, Bao C, Noble PW, Lane MD, Die AM. Leptin regulates
proinflammatory immune responses. FASEB J. 1998;Jan(1):57-65.

29. Sakane N, Yoshida T, Umekawa T, Kogure A, Kondo M. Beta 2-
adrenoceptor gene polymorphism and obesity. Lancet.
1999;June5;353(9168):1976.

30. Harding SM, Sontag SJ. Asthma and gastroesophageal reflux. Am J
Gastroenterol. 2000;Aug;95(8 suppl):823-32. Review.

31 . Dell S, To T. Breastfeeding and asthma in young children: findings from a
population-based study. Arch Pediatr Adolesc Med. 2001;Nov;155(11):1261-
5.

32.vonMutius E. Paediatric origins of adult lung disease. Thorax.
2001 ;Feb:56(2):153-7.

33.Annesi-Maesano l, Moreau D, Strachan D. ln-utero and perinatal
complications preceding asthma. Allergy. 2001;Jun;56(6):491-7.

34.Taylor WR, Newacheck PW. Impact of childhood asthma on health.
Pediatrics. 1992;90:657-662.

35. Munyard P, Bush A. How much coughing is normal? Arch Dis Child.
1996;June;74(6):531-4.

36. McKenzie S. Cough—but is it asthma? Arch Dis Child. 1994;Jan;70(1):1-2.
37.Montnemery P, Hansson L, Lanke J, Lindholm LH, Byberg P, Lodfahl CG,

Adelroth E. Accuracy of a first diagnosis of asthma in primary health care.
Fem Prac. 2002;Aug;19(4):365-8.

83

38. Flegal KM, Carroll MD, Kuczmarski RJ, Johnson CL. Overweight and obesity
in the United States: prevalence and trends, 1860-1004. Int J Obes Relat Med
Disord. 1998;22:39-47.

39.Troiano RP, Flegal KM, Kuczmarski RJ, et al. Overweight prevalence and
trends for children and adolescents: the National Health and Nutrition
Examination Surveys, 1963 to 1991. Arch PediatrAdolesc Med.
1995;149:1085-1091.

40.Wolf AM, Coldltz GA. Current estimates of the economic cost of obesity in the
United States. Obes Res. 1990;Mar 6(2):97 106.

41 .Wang G, Dietz WH. Economic burden of obesity in youths aged 6 to 17 years.
Pediatrics. 2002;May;109(5):E1-1 .

42.Spiegelman O, Israel RG, Bouchard C, Willett WC. Absolute fat mass,
percent body fat, and body-fat distributions: Which is the real determinant of
blood pressure and serum glucose? Am J Clin Nutr. 1992;55:1033-44.

43. Meyer HE, Sogaard AJ, Tverdal A, Selmer RM. Body mass index and
mortality; the influence of physical activity and smoking. Med Sci Sports
Exerc. 2000;July 34(7): 1 065-70.

44. Katzmarzyk Pt, Craig CL, Bouchard C. Adiposity, adipose tissue distribution
and mortality rates in the Canada Fitness Survey follow-up study. Int J Obes
Relat Metab Disord. 2002;Aug;26(8):1054-9.

45. Kalogeromitros D, Katsarou A, Annenaka M, Rigopoulos D, Zapanti M,
Stratigos I. Influence of the menstrual cycle on skin-prick test reactions to
histamine, morphine and allergen. Clin Exp Allergy. 1995 May;25(5):461-6.

46. Gauthier BM, Hickner JM, Omstein S. High prevalence of ovenrveight children
and adolescents in the Practice Partner Research Network. Arch Pediatr
Adolesc Med. 2000 Jun;154(6):625-8.

47.Stenilus Araniala B, Poussa T, Kvamstrom J, et al. Immediate and long-term
effects of weight reduction in obese people with asthma: randomized
controlled study. BMJ. 2000;320:827-33.

48. Chen Y, Horne SL, Dosman JA. Body weight and weight gain related to

pulmonary function decline in adults: a six-year followup study. Thorax.
1 993;Apr;48(4):375-80.

84

49. Kaplan BA, Brush G, Mascie-Taylor CGN. The relationship of childhood
asthma and wheezy bronchitis with height, weight and body mass index. Hum
Bio. 1987;59:921-31.

50.Gokbel H, Atas S. Exercise-induced bronchospasm in nonasthmatic obese
and nonobese boys. J Sports Med Phys Fitness. 1999 Dec;39(4):361-4.

51 . Dixon JB, O’Brien PE. Gastroesophageal reflux in obesity: the effect of lap-
band placement. Obes Surg. 1999;Dec 9(6):527-31.

52. Hakala K. Stenius-Aamiala B, Sovijarvi A. Effects of weight loss on peak flow
variability, ainrvays obstruction, and lung volumes in obese patients with
asthma. Chest. 200:1 18(5):1315-21 .

53. Macgregor AM, Greenberg RA. Effect of surgically induced weight loss on
asthma in the morbidly obese. Obes Surg. 1993 Feb;3(1 ):15-21.

54.Collins L, Hoberty P, Walker J, et al. The effect of body fat distribution on
pulmonary function tests. Chest. 1995;107:1298-1302.

55. Lazarus R, Sparrow D, Weiss ST. Effects of obesity and fat distribution on
ventilatory function. Chest. 1997;1 1 1 :891-8.

56. Ray CS, Sue DY, Bray G, Hansen JE, Wasserrnan K. Effects of obesity on
respiratory function. Am Rev Respir Dis. 1983;128:501 -6. g

57. lnselman LS, Milanese A. Deurloo A. Effect of obesity on pulmonary function
in children. Pediatr Pulmonol. 1993;16:130-7.

58. Nimmagadda SR, Evans R 3rd. Allergy: etiology and epidemiology. Pediatr
Rev. 1999 Apr;20(4):1 1 1-5.

59. Sporik R, Holgate ST, PIatts-Mills TA, Cogswell JJ. Exposure to house-dust
mite allergen (Der p l) and the development of asthma in childhood. A
prospective study. N Engl J Med. 1990 Aug 23;323(8):502—7.

60. Huang SL, Shiao G, Chou P. Association between body mass index and
allergy in teenage girls in Taiwan. Clin Exp Allergy. 1999;29(3):323-9.

61.Jedrychowski W, Maugeri U, Flak E, Mroz E, Bianchi l. Predisposition to
acute respiratory infections among overweight preadolescent children: an
epidemiologic study in Poland. Public Health. 1998 May;112(3):189-95.

62. Harding SM. Gastroesophageal reflux, asthma, and mechanisms of
interaction. Am J Med. 2001 Dec 3;111 Suppl 8A:8S-12$.

85

63.Andze GO, Brandt ML, St Vil D, Bensoussan AL, Blanchard H. Diagnosis and
treatment of gastroesophageal reflux in 500 children with respiratory
symptoms: the value of pH monitoring. J Pediatr Surg. 1991 Mar;26(3):295-9.

64.Bagnato GF, Gulli S, Giacobbe O, De Pasquale R, Purello D’ambrosio F.
Bronchial hyperresponsiveness in subjects with gastroesophageal reflux.
Respiration. 2000;67(5):507-9.

65.Barak N, Ehrenpreis ED, Harrison JR, Sitrin MD. Gastro-oesophageal reflux
disease in obesity: pathophysiological and therapeutic considerations. Obes
Rev. 2002 Feb;3(1):9-15.

66.Castell DO. Obesity and gastro-oesophageal reflux: is there a relationship?
EurJ Gastroenterol Hepatol. 1996 Jul;8(7):625-6.

67. Locke GR 3rd, Talley NJ, Fett SL, Zinsmeister AR, Melton LJ 3rd. Risk factors
associated with symptoms of gastroesophageal reflux. Am J Med. 1999
Jun;106(6):642-9.

68. Lagergren J, Bergstrom R, Nyren O. No relation between body mass and
gastro-oesophageal reflux symptoms in a Swedish population based study.
Gut. 2000 Jul;47(1):26-9.

69. Fisher BL, Pennathur A, Mutnick JL, Little AG. Obesity correlates with
gastroesophageal reflux. Dig Dis Sci. 1999 Nov;44(11):2290-4.

70.Gdalevich M, Mimouni D, Mimouni M. Breastfeeding and the risk of bronchial
asthma in childhood: a systematic review with meta-analysis of prospective
studies. J Pediatr. 2001;Aug;11139(2):261-6.

71 . Butte NF. The role of breastfeeding in obesity. Pediatr Clin North Am.
2001 :48(1):189-98.

72.von Kries R, Koletzko B, Sauerwald T, von Mutius E, Barnert D, Grunert V,
von Voss H. Breast feeding and obesity: cross sectional study. BMJ. Jul
17;319(7203):147-50. 147-50.

73.Gillman MW, Rifas-Shiman SL, Camargo Jr. CA, Berkey CS, Frazier AL,
Rockett HR, Field AE, Coldltz GA. Risk of overweight among adolescents
who were breastfed as infants. JAMA. 2001;285(19):2461-7.

74. Barr RG, Cooper DM, Speizer FE, Drazen JM, Camargo Jr. CA. Beta (2)-

adrenoceptor polymorphism and body mass index are associated with adult-
onset asthma in sedentary but not active women. Chest. 2001;120(5):1474-9.

86

75. Beckett WS, Jacobs Jr. DR, Yu X, lribarren C, Williams OD. Asthma is
associated with weight gain in females but not males, independent of physical
activity. Am J Respir Cn't Care Med. 2001;164(11):2045-50.

76.Turki J, Pak J, Green SA, Martin RJ, Liggett SB. Genetic polymorphisms of
the beta 2-adrenergic receptor in nocturnal and non-nocturnal asthma.
Evidence that Gly16 correlates with the nocturnal phenotype. J Clin Invest.
1995 Apr;95(4):1635-41.

77. Ekelund U, Aman J, Yngve A, Renman C, Westerterp K, Sjostrom M.
Physical activity but not energy expenditure is reduced in obese adolescents:
a case-control study. Am J Clin Nutr. 2002 Nov;76(5):935—41.

78.Vaz de Almeida MD, Graca P, Afonso C, D'Amicis A, Lappalainen R,
Damkjaer S. Physical activity levels and body weight in a nationally
representative sample in the European Union. Public Health Nutr. 1999
Mar;2(1A):105-13.

79.Chinn S, Rona RJ. Can the increase in body mass index explain the rising
trend in asthma in children? Thorax. 2001;56(11):845-50.

80.Anderson SD, Smith CD, Rodwell LT, Du Toit Jl, Robertson CF, Riedler J.
The use of nonisotonic aerosols for evaluating bronchial

hyperresponsiveness. In : Spector SL, editor. Provocative challenge
procedures. New York: Marcel Dekker; 1994:49-278.

81 . Busquets Monge RM, Call Combelles O, Checa Vizcaino MA, Garcia Algar O.
Epidemiological features of exercise-induced bronchial hyperresponsiveness
in children aged 13-14 years old in Barcelona (Spain). An Esp Pediatr.
2002;56(4):298-303.

87