Citation
Maternal exposure to carbon monoxide and risk of low birth weight among singleton births, between 1996 and 2005 in Denver, Colorado

Material Information

Title:
Maternal exposure to carbon monoxide and risk of low birth weight among singleton births, between 1996 and 2005 in Denver, Colorado
Creator:
Lee, Pei Chen
Publication Date:
Language:
English
Physical Description:
ix, 44 leaves : ; 28 cm

Subjects

Subjects / Keywords:
Carbon monoxide -- Toxicology -- Colorado -- Denver ( lcsh )
Birth weight, Low -- Colorado -- Denver ( lcsh )
Birth weight, Low ( fast )
Carbon monoxide -- Toxicology ( fast )
Colorado -- Denver ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 37-44).
General Note:
Department of Geography and Environmental Sciences
Statement of Responsibility:
Pei Chen Lee.

Record Information

Source Institution:
|University of Colorado Denver
Holding Location:
|Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
182519340 ( OCLC )
ocn182519340
Classification:
LD1193.L547 2007m L43 ( lcc )

Full Text
MATERNAL EXPOSURE TO CARBON MONOXIDE AND RISK OF
LOW BIRTH WEIGHT AMONG SINGLETON BIRTHS, BETWEEN
1996 AND 2005 IN DENVER, COLORADO
By
Pei Chen Lee
B.A., Fu-Jen Catholic University, 2004
A thesis submitted to the
University of Colorado at Denver and Health Sciences Center
In partial fulfillment
Of the requirements for the degree of
Master of Sciences
Environmental Sciences
2007


This thesis for the Master of Science
Degree by
Pei Chen Lee
Has been approved
by
Deborah Thomas
Rafael Moreno
Richard Miech
If7(0 hi-
Date


Lee, Pei Chen (M.S., Environmental Sciences)
Maternal exposure to carbon monoxide and risk of low birth weight among singleton
births, between 1996 and 2005 in Denver, Colorado
Thesis directed by Assistant Professor Deborah Thomas
ABSTRACT
Over the past decades, numerous studies have investigated the association
between ambient air pollution and adverse birth outcomes. This retrospect study
explored the association between the exposure of pregnant women to carbon
monoxide (CO) during different trimesters of pregnancy and the risk of lower birth
weight infants in Denver, Colorado. Birth certificates between 1996 and 2005 were
retrieved from Birth Registry maintained by Department of Public Health and
Environment of Colorado. The location of air monitoring stations, along with
monitoring data, was also obtained from the Colorado Department of Public Health
and Environment. All of these data were combined within a geographic information
system and the inverse distance weighted (IDW) interpolation method was used to
determine the concentrations of pollutants in the study area and estimate exposure
during each trimester of pregnancy. Multiple logistic regression models were used to
estimate the relative risk of lower birth weight in relation to maternal exposure levels
of individual CO air pollution. In this study, LBW was associated with maternal
exposure to high CO during the third trimester of pregnancy. Risk of LBW was also
found by per 1 ppm increase in CO concentrations in the third trimester (ORs 1.83;
95% Cl, 1.13-2.96).
This abstract accurately represents the content of the candidates thesis. I recommend
this publication.
Signed
Deborah Thomas


ACKNOWLEDGMENT
First, I would like to thank my parents for supporting me to complete my master
degree at University of Colorado at Denver and Health Sciences Center (UCDHSC).
Also, I would like to thank my thesis committee members for their guidance and
continuous assistance throughout this process.
Deb, I was fortunate to have the opportunity to study and learn with your guidance.
The time commitment involved with my thesis, the numerous revisions, discussions,
and personal attention are greatly appreciated.
Rafael, you were always helpful in responding to my GIS questions, respectful,
patient, and responded in a timely manner. Without your help aspects of my thesis
could not have been accomplished. Thank you.
Susan, your comments regarding my thesis were insightful and helpful. These
comments strengthened my thesis and for that I say thank you.
Richard, thank you for assisting me with my thesis. Your input and knowledge was
very helpful, given the fact that of the short time notice with this thesis. Tt was an
honor and a pleasure to have you on my thesis project.
I would also like to acknowledge and thank the following people and organizations
for their contributions: Dr. Beate Ritz from the Department of Epidemiology at
University of California at Los Angeles for giving some suggestions for the final
analysis; Dr. Chung Yi Li from the Department of Public Health at Fu Jen Catholic
University form his guidance for statistical methods; Mary, Kirk, and William of
Colorado Department of Public Health and Environment for providing me the birth
registry data and air pollution data; Andy of the writing center at UCDHSC; and
Brian my classmate for helping me go through my presentation.


TABLE OF CONTENTS
FIGURES.............................................................vii
TABLES.............................................................viii
CHAPTER
1. INTRODUCTION
Background..................................................1
The Problem.................................................2
Goals and Specific Aims.....................................3
2. REVIEW OF LITERATURE
Introduction................................................4
Biological mechanisms of CO from maternal smoking and LBW...4
LBW associations with CO air pollution......................6
LBW associations with indoor air pollution.................10
LBW non-associations to CO air pollution...................10
Exposure assessment for ambient air pollution..............13
Summary....................................................16
3. MATERIALS and METHODS
Subjects...................................................18
Exposure Assessment........................................19
Statistical Analyses.......................................23
4. RESULT
Demographic Characteristics
25


CO exposure and LBW..............................27
5. DISCUSSION..........................................32
6. CONCLUSIONS.........................................36
BIBLIOGRAPHY.............................................37
VI


LIST OF FIGURES
Figure
1. The location of CO air monitoring stations in Denver and neighbor counties ... 21
2. The CO concentration changes by different cell size........................22
3. The flow chart of performing IDW interpolation..............................22
4. Average concentrations for CO in different trimesters......................23
5. Mean birth weight, by month from 1996-2005..................................29
6. CO concentrations, by month from 1996-2005.................................29
vii


LIST OF TABLES
Table
1. Results of studies of ambient air pollution and low birth weight..............8
2. The effect period of air pollution and low birth weight in several published
studies.......................................................................9
3. Results of studies of non-association between ambient air pollution and low birth
weight.......................................................................12
4. Exposure assessment in several published studies............................14
5. The variables of each birth..................................................19
6. The addresses of CO air monitoring stations in Denver and neighbor counties...21
7. Variable description.........................................................24
8. Crude ORs (95% CIs) of potential confounding factors for term LBW
(N=27,839)...................................................................25
9. 25th, 50th, and 75th percentiles of CO concentrations of the first, second, and third
trimesters...................................................................27
10. Summary of birth weights for exposure to CO at different trimester........27
11. Odds ratios and 95% confidence intervals for term LBW by CO air pollution
viii


exposure: census block-level cohort in study areas.........................30
12. Association between birth weight and maternal exposure to CO in different of
trimester......................................................................31
IX


CHAPTER 1
INTRODUCTION
Background
Low birth weight (LBW, birth weight < 2500 grams) is one of the strongest
predictors of infant mortality and adverse health effects in later life (Joseph and
Kramer, 1996). Smaller birth size could be associated with biological, environmental,
lifestyle, and socioeconomic reasons (Institute of Medicine, 1985; Kramer, 1987 a, b;
Kramer et al., 1990; Silbergeld and Tonat, 1994). The first study that connected
maternal exposure to air pollutants and LBW was conducted in Los Angeles in the
1970s (Williams ct al., 1977). This study found that mothers who lived in areas with
high levels of air pollution during pregnancy had an increased risk of delivering LBW
infants. After that, many studies linked ambient air pollution to the incidence of LBW,
indicating that exposure to air pollutants such as carbon monoxide (CO), nitrogen
dioxide (NO2), sulfur dioxide (SO2), ozone (O3), particle diameter less than 10pm
(PM10), and particle diameter less then 2.5pm (PM2.5) in different trimesters may
increase the occurrence of low birth weight infants (Xu et al., 1995; Wang et al., 1997;
Pereira ct al., 1998; Bobak and Leon, 1999; Bobak, 2000; Wilhelm and Ritz, 2005;
Jennifer et al., 2005). Among those ambient air pollutants, CO air pollution has
shown to have greater effects on birth weight than the others (Rite and Yu, 1999; Ha
et al., 2001; Maisonet et al., 2001; Lee et al., 2003; Gouveia et al., 2004; Wilhelm and
Ritz, 2005). Also, a biological mechanism for fetal effects of CO has been presented
(Rite and Yu, 1999). Thus, the potential significant of CO pollution in regard to LBW
has become an important issue in environmental epidemiology. However, several
other studies found no such association (Alderman ct al., 1987; Chen ct al., 2002; Liu
et al., 2003; Parker et al., 2005), and so the nature of this association remains unclear.
Estimation of the concentration of exposure may be one of the reasons for
1


inconsistency of findings (Rogers and Dunlop, 2006).
Access to air monitoring data and birth registry data has made the study of
maternal exposure to ambient air pollution and birth defects possible. However, the
assessment of maternal exposure to air pollution has not been easy. The main reason
for this difficulty is that the amount of personal exposure through air monitoring data
alone is not known. The data cannot truly represent personal exposure concentration
since meteorological phenomena, such as wind direction and humidity, may change
the distribution of the pollution concentration. Also, unmeasured indoor pollution is
another challenge. Because of the methodological limitations of unsound exposure
assessment techniques, many studies incorporated certain techniques for improving
exposure assessment, such as geospatial interpolation methods (e.g. inverse distance
weighted and Kriging) and certain modeling techniques (the air pollution dispersion
model and environmental transport model).
In this research, I derived block-based environmental exposure estimates from air
monitoring station reports, following the methodology of Leem et al. (2006) that
examined preterm and air pollution in Incheon, Korea from January 1st, 2001 to 31
December, 2002. Instead of Kriging, this study used the inverse distance weighted
(IDW) statistical mapping technique to predict an average concentration of pollutants
in each census block because of the small number of CO monitoring stations in
Denver. This decision is justified by David et al. study (2004), in which they
suggested that The interpolation method depends strongly upon the nature of the
local monitor network. In regions of the country in which monitors are sparse, all
interpolation methods converge to a similar, narrow range of predictions. The
exposure estimates were combined with birth records at the block level within the
study area.
The Problem
Colorado has one of the highest LBW rates in the United States. In 1997, the
incidence rate of LBW was 8.9 percent in Colorado compared to 7.5 percent in the
2


United States (Colorado Department of Public Health & Environment, CDPHE,
2000). In 2003, the LBW rate in Colorado remained high at 9.1 percent (Niermeyer et
al., 2006). Inadequate maternal weight gain, smoking, premature (gestation week
smaller than 37 weeks), less maternal education, and altitude were factors identified
as contributing to LBW among singleton births (CDPHE, 2000). Among those risk
factors, inadequate maternal weight gain, smoking, and premature birth were
considered the top three leading causes of LBW. The first study of CO exposure and
risk of LBW in Denver found no association between higher CO exposure and higher
odds of LBW during the last 3 months of gestation (Alderman et al., 1987). However,
they were unable to adjust for maternal smoking and for the exposure
misclassification to CO because of local variations not reflected by data from
stationary monitors (Alderman et al., 1987).
Goals and Specific Aims
The primary goal of this study is to investigate the potential link between birth
weight and maternal CO exposure among singleton term births between 1996 and
2005 in the Denver area. The secondary goal of this study is to provide insight into
the relationship between CO air pollution and LBW in Denver, which may give
health care providers, policymakers, and pregnant women some directions for
reducing the LBW rate in Denver.
The specific aims of this study are:
Aiml: to assess CO air pollution as a risk factor for LBW
Aim2: to predict the CO concentration within a 2 mile buffering zone of the CO
monitoring stations
Aim3: to calculate the odds ratios (ORs) of LBW in relation to maternal exposure
to different quartiles of the concentrations of CO during each of the trimesters
3


CHAPTER 2
REVIEW OF LITERATURE
Introduction
Numerous studies have demonstrated that maternal smoking during pregnancy
increased the risk of delivering at LBW (Cnattingius et al., 1985; Cnattingius, 1989;
Cnattingius et al, 1993; Wen et al., 1990; Fox et al., 1994), and this effect has given
rise to the theory that maternal exposure to CO during pregnancy may increase the
risk of delivering at LBW (National Academy of Sciences, 1977; U.S. Environmental
Protection Agency, 1977; Longo, 1977). After this first hypothesis on ambient air
pollution, especially CO, and its deleterious effects on fetal growth and development
was established (National Academy of Sciences, 1977; U.S. Environmental Protection
Agency, 1977; Longo, 1977), many studies have examined the relationship between
ambient air pollution and LBW. Some of the studies showed a relationship between
ambient air pollution and LBW while others showed no such association. One of the
causes for these inconsistent findings might depend on how maternal exposure was
determined. Several of the assessments involved predicting pollutant concentration by
modeling or assigning the pollutant concentration from the nearest air monitoring
station to individual, zip-code, district, or county level. The following section will
describe in detail the relationship between biological mechanisms, maternal smoking,
and LBW; the relationship between ambient air pollution, particularly CO, and LBW;
and exposure assessment methods for ambient air pollution.
Biological mechanisms of CO from maternal smoking and LBW
The existence of biological mechanism for CO air pollution exposure and its
influence on LBW has not yet been completely established. However, there are some
experimental studies the effects of CO on fetal and newborn growth. In mammalian
4


studies, Wells exposed pregnant rats to 15,000 ppm CO for from 5 to 8 minutes
during the 21 day pregnancy and found that the surviving newborn rats failed to
growth normally (Wells, 1933). Astrup and co-workcrs exposed pregnant rabbits to
90 ppm CO for 30 days and found that the birth weights decreased from 57.7 to 51
grams and the neonatal mortality rate increased to 10 percent (Astrup et al., 1972).
Garvey and Longo exposed pregnant Lon-Evans rats to 30 or 90 ppm of CO and they
found that lung weight decreased 24 percent in those fetuses exposed to 90 ppm of
CO (Garvey and Longo, 1978). However, they did not find any changes in birth
weight among the fetuses exposed to CO.
In 1857, Claude Bernard first observed that carbon monoxide had greater affinity
for binding to hemoglobin (Hb) than oxygen. As a result, the oxygen tension of blood
decreases to lower than normal values. This effect may be particularly significant for
the fetus since the oxygen partial pressure in its arterial blood is normally relatively
low (about 20 to 30 mmHg) in the fetus as compared to adult values of about 100
mmHg (Behrman, 1971). In addition, fetal Hb has greater affinity for CO than
maternal Hb, the half-life of carboxyhemoglobin (COHb) in fetal blood is three times
longer than that of maternal blood, and the fetus has a higher rate of oxygen
consumption than mothers. Thus, maternal COHb concentration has a great influence
on the fetus COHb since CO crosses the placenta by simple diffusion and binds to
Hb forming COHb in the red blood cell stroma (Longo, 1977).
Maternal smoking affects fetal exposure to greater-than-normal CO concentrations.
In fact, many studies related maternal smoking and LBW (Simpson, 1957; Butler and
Alberman, 1969; Butler et al., 1972), infant mortality (Simpson, 1957; Frazier, et al.,
1961; Comstock and Lundin, 1967; Butler and Alberman, 1969; Fabia, 1973), and
body length (Miller and Hassanein, 1974). Among those studies relating to maternal
smoking and LBW, molecular and biochemical mechanisms have been found to show
a relationship between CO exposure from maternal smoking and a reduction of birth
weight that could provide some insight into the role CO might play (Gritz, 1980;
5


Davies et al, 1979; Everson et al, 1988; Wouters et al, 1987). Everson et al. (1988)
pointed out that maternal smoking during pregnancy is related to 3 of the DNA
adducts observed and the levels of adduction arc associated with caffeine use, age,
race, and the education of the mother (Everson et al, 1988). Thus, the results indicated
that there is a clear association between birth weight, birth length, and levels of
adduction among smoking mothers (Everson et al, 1988). The biochemical
mechanisms of maternal smoking, especially as related to the CO compound, and
LBW involve the binding of CO with hemoglobin to produce elevated levels of
HbCO in smokers, decreasing the oxygen available to the fetus (Davies et al, 1979).
In addition, CO exposure is combined with fetal blood flow reduction (Davies ct al,
1979). Wouters et al (1987) examined fetal outcomes and HbCO levels in the
umbilical cords of 77 uneventful pregnancies. They found that HbCO levels were
significantly elevated in the venous cord blood of children of smokers compared to
non-smokers (Wouters et al, 1987). This research provides evidence for biological
mechanisms for the negative impacts that smoking during pregnancy has on LBW.
LBW associations with CO air pollution
There is much epidemiological evidence for the link between ambient CO
exposure and the risk of LBW in different trimester of pregnancy. Table 1 shows the
results of studies regarding ambient air pollution and low birth weight. Two studies
conducted in Los Angeles reported a lower mean birth weight for babies whose
mothers lived in areas of higher concentrations of CO during the third trimester of
pregnancy (Ritz and Yu, 1999; Wilhelm and Ritz, 2005). In the Ritz and Yus study,
the birth cohort was bom between 1983 and 1993, including 125,573 singleton
children. The study found that maternal exposure to higher levels of ambient CO (>
5.5 ppm) during the last trimester of pregnancy was associated with significantly
increased risk of LBW [OR=T.22; 95% confidence interval (Cl), 1.02-1.44], The
second study included 146,972 infants bom between 1994 and 2000 in Los Angeles
and found the same association. This study indicated a 36% increase in LBW during
6


high CO exposure (>= 1.82 ppm) in the third-trimester pregnancy (Wilhelm and Ritz,
2005). However, the two studies were limited by their inability to control maternal
smoking and pregnancy wrcight gained, significant attributers to the risk of delivering
at LBW. Another study conducted in the United States examined the association
between ambient air pollution and LBW in the cities of Boston, Hartford,
Philadelphia, Pittsburgh, Springfield, and Washington (Maisonet et al., 2001). The
study found that maternal exposure to both CO and SO2 in the second and third
trimesters reduced the birth weight (Maisonet et al., 2001). Further, the study showed
that with every increase of lppm of CO exposure in the third trimester, the risk of
LBW increased (adjusted OR 1.31; 95% Cl, 1.06-1.62).
Internationally, two studies in Seoul, Korea also found that maternal exposure to
CO, SO2, and NO2 during the beginning or middle of pregnancy had adverse effects
on birth weight (Lee et al., 2003; Ha et al., 2001). The first study, completed by Ha et
al (2001), showed a small increase in the risk of LBW for each inter-quartile increase
of CO (adjusted relative risk 1.08; 95% Cl, 1.04-1.12) in the first trimester. Also, this
study indicated a negative relationship between birth weight and concentrations of
N02, SO2, and total suspended particle (TSP). This relationship was relatively linear,
without thresholds for concentrations of the pollutants (Ha et al., 2001). However,
maternal smoking was uncontrolled in these two Korea studies. Another study,
completed by Gouveia and colleagues, reported that a reduction of 23 g in birth
weight was estimated for a 1 ppm increase in mean exposure to CO during the first
trimester. Again, this study was unable to control maternal smoking and weight
gained during pregnancy.
Although much evidence indicates an association between CO ambient air
pollution and LBW, the exposure period is still an ongoing debate. Some studies
reported an association during the first trimester of pregnancy (Ha et al, 2001; Lee et
al, 20031; Gouveia et al, 2004), while others indicated the effect was in the third
trimester (Rite and Yu, 1999; Maisonet et al, 2001; Wilhelm and Ritz, 2005) (Table
7


2). Thus, more evidence is needed in order to be better establishing the effects of CO
exposure at various pregnancy periods on low birth weight.
Table 1. Results of studies of ambient air pollution and low birth weight
First Outcome and Exposure categories Odds ratio/changing
author exposure measures in birth weight (g)
(year) (95% Cl)
Rite LBW and mean Divided CO mean concentrations
(1999) concentrations of CO in into three categories (< 50th, 50 to <
third trimester of 95th, > 95th percentile)
pregnancy <50* vs. >95* percentile 1.22(1.03-1.44)
Ha LBW, and mean 1. Divided CO, 03, S02, N02, and TSP
(2001) concentrations of CO. mean concentrations into quartiles.
03, S02, N02, and TSP each quartile increased
in first and third CO 1.08(1.04-1.12)
trimesters of pregnancy so2 1.06(1.02-1.10)
no2 1.07(1.03-1.11)
TSP 2. Served pollutants as continuous 1.04(1.00-1.08)
variables
CO per 100 ppb increased 11.55(8.99-14.10)
S02 per 1 ppb increased 8.06(5.59-10.53)
N02 per 1 ppb increased 8.41(6.07-10.76)
TSP per 50.0 pg/m3 increased 6.06(3.85-8.27)
Maisonet LBW and mean 1. Divided CO, PMio, and S02 mean
(2001) concentrations of CO, concentrations into five categories
PMio, and S02 in each (< 25th, 25th to < 75th, 75th to < 95th,
trimester of pregnancy > 95th percentile) < 25th vs > 95th percentile
CO 1.15(0.94-1.42)
so2 2. Served pollutants as continuous 1.06(0.76-1.47)
variables
CO per 1 ppm increase 1.31(1.06-1.62)
Lee LBW and mean Divided CO, 03, S02, N02, and
(2003) concentrations of CO, PMIO mean concentrations into
S02, N02, and PM10 quartile for each quartile increase
in each trimester and CO 1.05(1.01-1.09)
month of pregnancy S02 1.14(1.04-1.24)
N02 1.04(1.00-1.08)
PMIO 1.06(1.01-1.10)
8


Table l(Cont). Results of studies of ambient air pollution and low birth weight
First author (year) Outcome and exposure measures Exposure categories Odds ratio/ changing in birth weight (g) (95% Cl)
Gouveia LBW and mean 1.Divided PMi0 S02, and 03 mean
(2004) concentrations of PM10 concentrations into quartiles (< 25th, CO, S02, N02 and 03 25th to 50th, 51st to 75th, > 75th in each trimester of percentile)
pregnancy <25 vs >75 percentile PM10 CO 2. Served pollutants as continuous variables CO per 1.0 ppm increase PMio per 10.0 ug/m3 increase 1.14(0.87-1.49) 1.02(0.82-1.27) -23.1(-41.3 to -4.9) -13.7(-27.0to -0.4)
Wilhelm LBW, preterm births, 1. Divided CO, PM10, and PM2 5 mean
(2005) and mean concentrations of CO, PM10, and PM2 5, in third trimester of fvrQon L/i vtuluiivj concentrations into three categories (< 25th, 25 to < 75*, and > 75* percentile) 2. Served pollutants as continuous variables CO per 1 ppm increase PMio >45.1 ug/m3 1.12(1.05-1.19) 1.12(0.91-1.38)
infants who are below the 10th percentile of birth weight for gestation week
Table 2. Effect periods of air pollution and low birth weight in several published
studies
Study period (First author) Area Air Pollutant Effect period
1989-1993(Rite) LA, Cahfomia CO Third trimester
1996-1997(Ha) Seoul, Korea CO, so2, no2, tsp First trimester
1994-1996(Maisonet) Northeastern United States co, so2 Second and thud trimesters
1996-1998(Lee) Seoul, Korea co, so2, no2, PM,o Early to mid pregnancy
1997(Gouveia) San Paulo, Brazil PM,o, CO First trimester
1994-2000( Wilhelm) LA, California PMio, CO Third trimester
9


LBW associations with indoor air pollution
In developing countries, up to 90% of rural households rely on biomass fuels in the
form of wood, dung, and crop residues for cooking and space heating (World
Resources Institute, 1998). Smoke from biomass combustion produces a large number
of substances which can damage human health, such as particles, carbon monoxide,
nitrous oxides, sulphur oxides (principally from coal), formaldehyde, and polycyclic
organic matter (De Koning et al., 1985). Women and young children are exposed to
high levels of indoor air pollution every day since women involved in cooking and
young children are often carried on their mothers (Bruce et al., 2000). According to a
survey in two districts of Zimbabwe in 1988, Bchcra ct al. found that on average
women spend 5 hours per day in the kitchen area, and the levels of CO in the kitchen
were in the range of 300 to 1000 ppm (Behera et al, 1988). There are many studies
that have examined health effects of biomass smoke (Mavalankar et al., 1991; Behera
et al, 1988; and Boy et al., 2002 ), including associations with low birth weight. Boy
et al. (2002) reported an association between household use of wood fuels and
reduced birth weight (Boy et al., 2002). This study found that babies bom to mothers
using wood fuels were 63 grams lighter than those bom to mothers using gas or
electricity (Boy et al., 2002). Another study conducted in Zimbabwe also found that
babies bom to mothers cooking with wood, dung, or straw were 175 g lighter than
those bom to mothers using liquid petroleum gas (Mishea et al., 2004).
LBW non-associations to CO air pollution
Several studies have refuted an association between CO air pollution and LBW.
Alderman et al. (1987) was the first study examining air pollution and LBW in
Denver, Colorado. This study did not find an association between higher CO
exposure and higher risk of LBW. The bias in this study was that the local variation
in CO concentration was not incorporated from stationary monitors because this
variation was very small (Alderman et al., 1987). They also did not control for
maternal smoking. Thus, they recommended that further studies would likely require
10


direct measurement of total CO exposure, which includes indoor and workplace CO
exposure. Another study conducted in California found that higher PM2.5 exposure
increased the risk of small for gestation age (SGA) (Parker ct al, 2005). However,
they did not find a link between CO exposure and LBW. This finding varied from the
two other studies described previously conducted in California (Ritz and Yu, 1999;
Wilhelm and Ritz, 2005). The disparity in results could possibly be because of
differences in the measures of CO exposure. Another study in Northern Nevada was
also unable to find a link between CO exposure and LBW (Chen et al., 2002). This
study examined 39,338 singleton births in Washoe County from 1991 through 1999.
However, CO was not found to be associated with birth weight through logistic
regression after controls for infant sex, city of residence, education, medical risk
factors, active tobacco use, drug use, alcohol use, prenatal care, age, race, ethnicity,
and maternal weight gain. In Vancouver, Canada, non-association between CO
exposure, and LBW was found (Liu et al., 2003). However, preterm birth was
associated with CO during the last trimester of pregnancy and the odd ratios for 1
ppm increase of CO was 1.08 (95 % Cl, 1.01-1.15). Inability to control maternal
smoking was the common limitation for many of these studies (Alderman ct al., 1987;
Chen et al., 2002; Parker et al, 2005) (Table 3).
11


Table 3. Results of studies of non-association between ambient air pollution and low
birth weight
First Outcome and Non-association Limitations
author (year) exposure measures
Alderman LBW and mean No statistically significant 1. Potential source of bias in
(1987) concentration of CO effects were noted of CO. the measurement of CO
during the last 3 months exposure
of pregnancy 2. The range of CO exposures available
3. Over adjustment for the effects of extraneous variables
4. Inability to adjust maternal smoking
Chen LBW and mean 1. No associations 1. Inability to adjust fathers
(2002) concentrations of CO, between exposure to race and ethnicity, marital
PMio, and 03 in each ambient concentrations status, environmental
trimester of pregnancy of CO, 03 and term tobacco exposure,
LBW occupational exposures,
2. PM10 was found to be and active smoking
related to the risk of history
LBW from multiple 2. Assumption that personal
linear regression. exposure to air pollutants
However, the was equal to ambient air
association was not found from logistic regression. pollution levels
Liu LBW, preterm births, No associations between 1. Inability to control
(2003) intrauterine growth exposure to ambient socioeconomic status
retardation (IUGR), and concentrations of CO, 2. The misclassification of
mean concentrations of N02, O3, and term LBW unknown individual
CO, SOj, N02 and 03, in each trimester of maternal exposure
pregnancy
Parker LBW, small for No associations between 1. Deciding on the
(2005) gestational age (SGA)a, exposure to ambient appropriate measurement
and mean concentrations of CO and of pollution
concentrations of CO term LBW 2. The effect of geographic
and PM2.5 in each unit of exposure
trimester of pregnancy measurements is unclear
3. Inability to adjust maternal smoking
12


Exposure assessment for ambient air pollution
Assessment of maternal exposure to air pollution has many challenges. Most of
the previous studies used zip-code-level analysis, incorporating air monitoring station
data or constant distance from the air monitor to determine levels of exposure for
mothers (Wang et al., 1997; Bobak and Leon, 1999; Bobak, 1991; Wilhelm and Ritz,
2005). Such methods, however, may entail certain degrees of exposure
misclassification bias and such bias could ultimately lead to an overestimation or
underestimation of the true effects (Brenner et al., 1992; Morgenstem and Thomas,
1993). A few studies estimated individual exposure to ambient air pollution by
modeling from air monitoring station to a geographically identified receptor, through
the use of an environmental transport model, or block Kriging statistical mapping
technique, which can predict an average concentration over a spatial region from
point locations (Rogers and Dunlop, 2006; Leem et al., 2006).These methods, which
specified the individualized environmental exposures, largely reduce the
misclassification bias. Therefore, they are considered better than the zip-code-level
analysis and constant distance from the air monitoring station.
Several attempts have been made to associate CO with LBW (Rite and Yu, 1999;
Ha et al., 2001; Maisonet et al., 2001; Lee et al., 2003; Gouveia et al., 2004; Wilhelm
and Ritz, 2005; Alderman et al., 1987; Chen et al., 2002; Liu et al., 2003; Parker et al.,
2005). Two studies conducted in Denver and Los Angeles (Alderman et al., 1987; and
Rite and Yu, 1999) used mean CO concentrations measured from air monitoring
stations and examined the mothers who were residing in the census block or zip-code
area within a 2-mile radius of stations to examine the effect on birth weight (Table 4).
The median exposure levels reported for Denver were ranged from 0.5 to 3.6 ppm,
and the 50-95-percentile exposure levels for Los Angeles were 2.2 to 5.5 ppm.
Wilhelm and Ritz (2005) estimated exposure levels by calculating the distance to the
nearest air monitors, and they suggested that the effects manifested during the first
and second trimester as the result of CO concentrations for mothers who residing
13


within a 1 mile distance of monitors. Another series of studies that examined the CO
effect on birth weight also relied on ambient measurements of the CO air pollutant for
exposure estimates (Ha ct al., 2001; Maisonct ct a!., 2001; Lee ct al., 2003; Gouvcia
et al., 2004; Alderman et al., 1987; Chen et al., 2002; Liu et al., 2003). These studies
assigned mean concentrations of different ambient air pollutants to mothers who lived
within the same cities as the stations. Mean exposure levels within each trimester
were calculated by averaging daily ambient air pollution concentrations during the
corresponding days.
Tabic 4. Exposure assessment in several published studies
First author (year) Exposure assessment Analysis level
Alderman (1987) 2 mile radius of CO monitoring stations census-tract level
Rite (1999) 2 mile radius of CO monitoring stations zip-code level
Ha (2001) mothers living within the same city as monitoring stations city level
Maisonet (2001) mothers living within the same city as monitoring stations city level
Chen (2002) mothers living within the same city as monitoring stations city level
Lee(2003) mothers living within the same city as monitoring stations city level
Liu (2003) link environmental air pollution data at the census subdivision level to birth records census-subdivision level
Gouveia (2004) assigned the nearest air monitoring stations data to each mother city level
Wilhelm (2005) 1 mile, 1 to 2 mile, and 2 to 4 mile radius of CO, PM]0, and PM25 monitoring stations zip-codc level personal level
Parker (2005) 5 mile radius of CO and PM2 5 monitoring stations personal level
Due to data availability and privacy issues, individual geographic location is not
easily obtainable from the birth registry dataset. However, two studies conducted in
California were able to obtain personal levels of information such as coordinates and
addresses to identify the residence of mothers (Wilhelm and Ritz, 2005; Parker ct al.,
14


2005). In addition, both studies were able to calculate the exact distance from
maternal residence to the nearest air monitoring stations. Although the studies were
capable of estimating the distance to each of the stations, personal exposure
concentration still could not be obtained from monitoring data.
Among modeling techniques, the environmental transport model, which is based
on the Gaussian plume atmospheric transport model, was used by Rogers and Dunlop
(2006) to estimate PMio exposure at the scale of the birth home. The annual PM]0
emissions were measured from air monitoring stations and industrial point sources.
The Gaussian plume atmospheric transport models assume that plume concentration
has independent Gaussian distributions in horizontal and vertical planes at each
downwind distance (Holland, 1953; Meade and Pasquill, 1951; Hamby, 2002). This
model is widely used to predict concentrations in the atmosphere and is recommended
by the US EPA (Hanna and Briggs, 1982). Although the environmental transport
model had been validated elsewhere (Rogers et al., 1999), the primary limitation
involves the uncertainties in the model, such as release quantities, the transport model,
ambient temperature, wind speed and direction, and stack gas temperature and
velocities (Rogers and Dunlop, 2006).
As another interpolation method, Kriging is a weighted average technique used
to estimate a smooth surface from data points over the domain and predict the average
block based on regular grids. Additionally, cross-validation, a technique with which
each monitoring station is removed, one at a time, and the concentration at each
omitted station is predicted using the concentration values observed at the other
monitors, is used to evaluate the quality of the predicted values from Kriging. The
first study to use the Kriging statistical mapping technique was presented by Lccm ct
al. (2006). They used 0.17 km 0.17 km grids to partition each dong (the
administrative unit in Korea similar to a county) for each pollutant and each month.
They found the Kriging technique provided reasonable results for surface
interpolation of pollutants concentrations, such as CO, SO2, NO2, and PMio. Because
15


this method is fairly accurate and avoids the artifacts that often result from the use of
inverse distance weighted, spline, or global/local polynomials, its use for predicting
the spatial distribution of air pollutants was one of the strengths of this study (Jcrrctt
et al., 2005; Ritz et al., 2000; Waller and Gotway, 2004). However, they still were not
able to control for maternal smoking, its potential misclassification of exposure due to
the use of surrogate ambient air pollution data, the uncertainty of the predicted
average concentrations for the dongs, and its inability to geocode the residential
addresses to point locations.
Summary
The fast expansion of studies examining the relationship between maternal
exposure to ambient air pollution and adverse pregnancy outcomes, especially LBW,
was enabled by the existence of air monitoring stations and routinely collected birth
certificate information in many urban areas. The studies that examined the association
between CO exposure and LBW show that the ambient CO air pollutant seems to play
some role in determining birth weight. However, the differences in study outcomes
and pregnancy periods studied result in the studies undergoing continued debate
(Wilhelm and Ritz, 2005). Reduction of birth weight varies according to the timing of
exposure (Gouveia et al., 2004). The possible biological reasons for interference with
the final birth weight are the formation of the placenta occurring during first trimester
and weight gain occurring predominantly during the third trimester (Mongelli and
Biswas, 2001).
Estimating environmental exposure to ambient CO pollution is another challenge
in epidemiology studies. Most of the studies rely on crude methodologies obtained
from air quality monitoring data by sampling networks and assigning a crude estimate
of exposure in large population areas such as cities, counties, census blocks, and
postal zip code areas (Rogers et al., 1999). These kinds of ecological studies ignore
the location of pollutant sources and the atmospheric influence on concentrations of
pollutants. As a result, these ecological studies cannot identify the mothers as exposed
16


or unexposed. Recently, many studies have incorporated air modeling techniques,
such as the environmental transport model and/or GIS spatial modeling techniques to
improve the exposure assessment method and to reduce the misclassifieation in
previous studies, which assigned a crude estimate of exposure to large population.
17


CHAPTER 3
MATERIALS and METHODS
Subjects
Birth records were obtained from the Colorado Birth Registry database
maintained by the Colorado Department of Public Health and Environment. The
information for each live birth included: demographic information, geographic
information, maternal pregnancy history and some risk factors of LBW (Table 5).
Due to the privacy and confidentiality of individual health birth records, the home
address location for each birth was not available, but instead was obtained at the
census block level. Thus, the scale of analysis was at a block level because of the way
in which the data could be released. From this dataset, all live singleton births bom at
normal gestation weeks (between 37- 42 weeks of gestation) during the period
September 1st, 1996 to December 31,2005 were selected. In addition to eliminating
multiple births and births outside of the normal gestation period, live births by birth
weights greater than 5000 grams (g) or smaller than lOOOg were also excluded. Based
on previous studies, this exclusion should not alter findings (Ritz and Yu, 1999).
296,593 singleton term births were bom in Denver during the study periods. 3,770
blocks were selected in this study and resulted in a total of 27,839 births during the
study periods. Some subjects were excluded because of a missing census block
identifier (n=3,846) and census blocks outside of the study area (n=264,908). Also,
some study subjects may have been excluded in our adjusted analyses because of
missing data for covariates such as infant sex, married status, maternal race, maternal
education, gestation week, cigarette smoking during pregnancy, type of last delivery,
season of birth, prenatal visits, and weight gained during pregnancy.
18


Table 5. The variables of each birth
Variable Name Purpose
Certificate Number To identify each births
Year of Birth To identify births between 1996 and 2005
Month of Birth To calculate the trimester of pregnancy To assign the monthly CO concentration To define the season of birth
Day of Birth To calculate the trimester of pregnancy
Sex The adjust factors
Age of Mother The adjust factors
Race of Mother Code The adjust factors
Education of Mother The adjust factors
Mother Married The adjust factors
Birth Weight Dependent outcome
Type of Last Delivery To identify the pregnancy history
Clinical Estimate of Gestation To identify term births The adjust factors
Cigarettes per day The adjust factors
Weight Gained The adjust factors
Prenatal Visits The adjust factors
Previous Preterm or Small-for- Gestational Age Infant Medical risk factors
Census Block Geographic Variable As a key variable merge to pollution data
Exposure Assessment
Maternal exposure to CO air pollution during various pregnancy periods was
estimated based on air monitoring data maintained by the Air Pollution Division of
the Colorado Department of Public Health and Environment. CO pollutant data were
routinely collected from 4 monitoring stations in Denver and 3 monitoring stations
located in the Denver metropolitan areas (Figure 1, and Table 6). These monitoring
stations were generally located around industrial and residential areas. For each of the
CO stations, the data included 24 one-hour concentrations. In this study, the subjects
monthly exposure to CO air pollination was determined from the CO concentrations
at the monitoring stations by using the block IDW statistical mapping technique,
which was performed with lire Geostatistical Analyst extension of AreGIS (AreMap,
version 9.1; ESRI Inc., Redlands, WA, USA). A 40m 40m size grid was used in this
19


study. The CO concentration by different cell size was tested and it was found that the
CO concentration at the cell size of 40 m had little variation when compared with
other cell sizes below 100 m (Figure 2). In addition, there were at least 12 grids
within each census blocks, since the mean area of each census block in the study
areas is 19,704 m2.
The study area was defined within a 2 mile buffering zone of the CO monitoring
stations. In addition, some boundary census blocks were kept if the block areas fell at
least 60% within a 2 mile buffering zone of the CO monitoring stations. The 2-mile
criterion is based on the previous study conducted in Los Angeles, USA (Wilhelm
and Ritz, 2005). They suggested that stationary' air monitors may most accurately
reflect CO air pollution within a small area surrounding stations because CO
concentrations vary spatially according to local sources (Wilhelm and Ritz, 2005).
Monthly average concentration for CO was calculated for each site. Based on
those concentrations, the IDW method was utilized to predict the monthly
concentrations of CO in the study area. After that, monthly concentrations of CO in
each census block were determined by performing Zonal Statistics function in
ArcMap (Figure 3). Referring to Lccms study (2006), the monthly CO
concentrations were assigned according to the gestation age, year and month of
delivery of each mother (Figure 4). The CO exposure for each trimester was
calculated by averaging monthly pollutant data into a summary exposure measure. In
an ideal situation, the study would start on the end of the month, but this is rarely the
case. For example, if a mother was delivering in mid-October and the gestation age
was 40, the pregnancy period would be divided into three trimesters. These periods
arc: mid-January' through mid-March, mid-April through mid-June, mid-July through
delivery month. Then, the average exposure concentration for CO pollutant in
different trimesters was calculated by averaging monthly pollutant data. Thus, the CO
concentration for the first trimester in this example would be the average
concentrations of mid-January through mid-March.
20


Figure 1. The location of CO air monitoring stations in Denver and neighbor counties
Table 6. The addresses of CO air monitoring stations in Denver and neighbor counties
County City Zip Address
Adams Welby 80229 3174 E. 78th Ave
Arapahoe Littleton 80122 8100 S. University Blvd
Denver Denver 80205 2105 Broadway
Denver Denver 80220 14th Ave & Albion St.
Denver Denver 80211 2325 Irving St.
Denver Denver 80204 1300 Blake St.
Jefferson Arvada 80002 9101 W. 57th Ave
21


Figure 2. The CO concentration changes by different cell size (The coordinates for
the point are 495431.360,4402565.053.)
Figure 3. The flow chart of performing IDW interpolation
22


Calculate the date of conceived
(date of delivering gestation weeks)
I
Calculate monthly average concentrations for CO
I
Temporary assigned monthly CO concentration to gestation month
January February March April May June July August September October
(First Trimester) (Second Trimester) (Third Trimester)
I________________________ I

Calculate average CO concentrations for each trimester
1
Calculate ORs
Figure 4. Average concentrations for CO in different trimesters
Statistical Analyses
The association between CO air pollution and LBW was evaluated using logistic
regression. LBW was served as the dependent variable and the quartiles of CO
concentration values were used to assign relative exposure categories (< 25th, 25th to <
75th, 75th to < 95th, and >= 95th percentiles). From previous studies, the pollutant
concentration was incorporated as categories and quartiles for the distribution of the
concentration values for each pollutant (Leem et al., 2006; Ritz et al., 2002; Wilhelm
and Ritz, 2003; Jennifer et al., 2005; Dugandzic et al., 2006; Gouveia et al., 2004). In
this study, the CO concentrations in 25th and 50th are close and so the CO
concentrations between 25th percentile and below 75th percentile were grouped as the
same category. The ORs calculated from the multiple logistic regression were used to
estimate the relative risk associated with maternal exposure to CO air pollution,
individually with simultaneous adjustment for infant sex, gestation week, married
23


status, maternal age, maternal race (Hispanic, White, Black, Others), maternal
education, cigarette smoking during pregnancy, conception season, prenatal care, and
weight gained during pregnancy (Table 7). Based on previous studies, these variables
consider the risk factors for delivering LBW, especially maternal smoking.
Additionally, LBW was incorporated into another model as continuous data in order to
examine whether there was a dose-response relationship between maternal exposure
to CO and risk LBW in different trimesters. Risk factors for LBW that are not
registered on Denver birth certificates include: maternal passive smoking and
maternal weight and height. In addition, this study cannot examine the CO interaction
with other ambient air pollutants because of spare air monitoring stations for other
pollutants in Denver.
Table 7. Variable description
Dependent Variable? Independent Variables Statistical model
LBW CO concentration
Grouped them into categories < 25 25th to < 75th, 75 to < 95th, and > Logistic regression
95th percentile Served them as continuous variable Linear regression
Controlling factors: infant sex, gestation week, maternal age, married status, maternal race, maternal education, cigarette smoking during pregnancy, conception season, prenatal care, and weight gained during pregnancy
24


CHAPTER 4
RESULTS
Demographic Characteristics
The mean birth weight for term LBW infants was 2,321 g compared to mean of
3,326 g for normal birth weight infants. Crude ORs with 95% Cl for the potential
confounding factors for LBW are shown in Table 8. Among those risk factors, the
ORs for LBW tended to be higher among infants whose mothers were education
between 9 to 15 years, unmarried, smoking, or inadequate weight gained (> 30
pounds) during pregnancy. The ORs was also higher among infants whose mother did
not have prenatal care. In female infants, the ORs of LBW was higher than in male
infants.
Table 8. Crude ORs (95% Cl) of potential confounding factors for term LBW
(N=27,839)
Characteristics Term LBW n=972 Normal Weight n=26,867* Crude OR 95% Cl
n % n %
Maternal age (year) <20 173 17.8 3,922 14.6 1.20(0.95-1.51)
20-24 279 28.7 6,829 25.4 1.11(0.90-1.37)
25-29 220 22.6 6,303 23.5 0.95(0.76-1.18)
30-34 168 17.3 6215 23.1 0.74(0.58-0.93)
>35 Maternal education 132 13.6 3594 13.4 1
(years) 6-8 96 10.1 3,619 13.6 1.11(0.86-1.43)
9-11 305 31.9 6,746 25.4 1.89(1.57-2.28)
12 267 28.0 5,722 21.6 1.95(1.61-2.37)
13-15 110 11.5 3,052 11.5 1.51(1.18-1.92)
16+ 177 18.5 7,396 27.9 1
25


Table 8(Cont). Crude OR (95% Cl) of potential confounding factors for term LBW
(N=27,839)
Characteristics Term LBW n=9724 Normal Weight n=26,867 * Crude OR 95% Cl
n % n %
Maternal race
Hispanic 18 1.9 358 1.3 1
White 750 77.2 23,358 86.9 0.64(0.40-1.03)
Black 176 18.1 2,608 9.7 1.34(0.82-2.211
Other 28 2.9 541 2.0 1.03(0.56-1.89)
Marital status
Married 481 49.5 17,017 63.3 1
Unmarried 491 50.5 9,850 36.7 1.76(1.55-2.01)
Infant sex
Male 396 40.7 13,696 51.0 1
Female 576 59.3 13,171 49.0 1.51(1.33-1.72)
Type of Last Delivery
Live 443 80.3 12,683 81.0 1
Fetal death 109 19.7 2,976 19.0 1.05(0.85-1.30)
Place of birth
Hospital 961 98.9 26,693 99.4 1
Residence 9 0.9 164 0.6 1.52(0.78-2.99)
Other 2 0.2 10 0.0 5.56(1.22-25.39)
Prenatal Care
Yes 893 95.4 25,681 98.6 1
No 43 4.6 366 1.4 3.38(2.45-4.67)
Weight gamed during Pregnancy (pounds)
<20 322 36.3 5,645 22.3 2.82(2.21-3.59)
21-30 323 36.4 8,410 33.3 1.90(1.49-2.42)
31-40 159 17.9 7,058 27.9 1.11(0.85-1.45)
41 or more 84 9.5 4,145 16.4 1
Cigarette smoking during Pregnancy
Yes 198 20.4 2,284 8.5 2.75(2.34-3.24)
No 774 79.6 24,583 91.5 1
Conception Season
Spring 256 n £ 96.3 1.09(0.91-1.30)
Summer 225 3.3 6,672 96.7 0.95(0.79-1.14)
Fall 243 3.5 6,603 96.5 1.03(0.86-1.24)
Winter 248 3.4 6,967 96.6 1
The calculation was based on insufficient number of cases due to the missing values.
26


CO exposure and LBW
The mean CO concentrations are 0.876,0.856, and 0.837 for the first, second, and
third trimester, respectively. Table 9 represents the different percentiles of CO
concentrations during each trimester. The CO concentrations in each percentile are
slightly higher in the first trimester than in second and third trimester, and the
concentrations in each percentile are relative lower in third trimester than in first and
second trimester.
Table 9. 25th, 50th, 75th, and 95th percentiles of CO concentrations of the first, second,
and third trimesters
CO Concentrations (ppm)
25 50th 75th 95th
First trimester 0.652 0.842 1.052 1.411
Second trimester 0.627 0.823 1.027 1.400
Third trimester 0.613 0.799 1.010 1.366
A summary of birth weight and percentage of LBW for exposure to CO at
different trimester is given in Table 10. The differences in birth weight and percentage
of LBW for different exposure-level groups were tested and no differences were
found.
Table 10. Summary of birth weights for exposure to CO at different trimester
Percentile of average CO exposure (ppm) Number of births Mean Birth Weight Standard Deviation (SD) Number of LBW Percent of LBW
First trimester < 0.652 6,951 3295.1 437.0 242 3.5
0.652 to < 1.052 13,934 3291.3 441.6 477 3.4
1.052 to < 1.411 5,558 3289.6 444.3 207 3.7
1.411 > 1,396 3269.1 439.6 46 3.3
27


Table 10(Cont). Summary of birth weights for exposure to CO at different trimester
Second trimester
< 0.627 6,957 3290.6 440.3 245 3.5
0.627 to < 1.027 13,908 3286.6 439.6 491 3.5
1.027 to < 1.400 5,583 3298.0 443.3 191 3.4
1.400 s Third trimester 1,391 3304.7 447.3 45 3.2
<0.613 6,966 3290.3 436.7 234 3.4
0.613 to < 1.010 13,910 3293.7 443.2 479 3.4
1.010 to < 1.366 5,569 3289.0 436.0 196 3.5
1.366 > 1,394 3272.4 458.1 63 4.5
In Figures 5 and 6, the present the mean birth weight and mean CO concentrations
are presented by month from September 1996 to December 2005. The mean birth
weight is predominantly between 3,250 g to 3,350 g during this study period.
However, the mean birth weight is about 3,072 g in September 1996. The potential
reason for the data discrepancy during September 1996 is that this only looked at 37-
39 gestation weeks, while the rest of the periods looked at 37-42 gestation weeks. The
CO monthly mean concentration diminishes by year and also exhibits a seasonal trend
with the concentration higher during winter time.
28


do >V <\ ^ ^ % r$> fe v<> S month
Figure 5. Mean birth weight, by month from 1996-2005
^ ^ month
Figure 6. CO concentrations, by month from 1996-2005
29


Table 11 represents the ORs for CO exposure and LBW after adjusting for infant
sex, gestation week, maternal age, married status, maternal race, maternal education,
cigarette smoking during pregnancy, conception season, prenatal care, and weight
gained during pregnancy. Crude ORs for LBW in relation to exposure to above 95th
percentile CO concentration was significantly increased in the third trimester in this
study. Adjusted ORs for LBW in relation to exposure to high (above 75th percentile)
CO concentrations were also significantly increased in the third trimester. There was a
significant association between LBW and high CO (above 75th percentile) exposures
in third trimester of pregnancy. Risk of LBW was not found by per 1 ppm increase in
CO concentrations in each trimester. Estimated coefficients from linear regression
models of birth weight as a continuous variable still cannot find any changes in mean
birth weight (Table 12).
Table 11. Odds ratios and 95% confidence intervals for term LBW by CO air
pollution exposure: census block-level cohort in study areas8
Total no. of singleton births Total no. of LBW Crude ORs (95% Cl) Adjusted ORsb (95% Cl)
First trimester Continuous 26,867 972 1.06(0.85-1.32) 1.03(0.63-1.67)
(per 1 ppm) < 0.652 6,951 242 1.00 1.00
0.652 to < 1.052 w 477 0.97(0.82-1.16) 0.94(0.75-1.18)
1.052 to < 1.411 6,991 207 1.12(0.91-1.39) 1.04(0.75-1.18)
1.411 > 6,954 46 1.01(0.72-1.41) 0.78(0.49-1.26)
Second trimester Continuous 26,867 972 0.97(0.78-1.21) 0.57(0.33-0.99)
(per 1 ppm) < 0.627 6,957 245 1.00 1.00
0.627 to < 1.027 6,945 491 0.94(0.79-1.12) 0.92(0.74-1.15)
1.027 to < 1.400 6,963 191 0.86(0.69-1.09) 0.75(0.53-1.05)
1.400 > 6,974 45 0.79(0.55-1.14) 0.68(0.41-1.14)
30


Table 11 (Cont). Odds ratios and 95% confidence intervals for term LBW by CO air
pollution exposure: census block-level cohort in study areas8_____________________
Third trimester Continuous 26,867 972 1.20(0.96-1.51) 1.83(1.13-2.96)
(per 1 ppm) <0.613 6,966 234 1.00 1.00
0.613 to < 1.010 6,956 479 1.03(0.87-1.22) 1.23(0.98-1.54)
1.010 to <1.366 6,954 196 1.11(0.90-1.36) 1.43(1.04-1.97)
1.366 > 6,963 63 1.47(1.08-2.00) 1.67(1.05-2.66)
a Includes the census block which fell at least 60% of area within a 2 mile buffering zone of
CO monitoring stations.
b Adjusted for infant sex, gestation week, maternal age, married status, maternal race,
maternal education, cigarette smoking during pregnancy, conception season, prenatal care,
and weight gained during pregnancy.
Table 12. Association between birth weight and maternal exposure to CO in different
of trimester
Birth Weieht. mean e Crude Adjusted8
CO Concentrations(ppm) 13 95% Cl 13 95% Cl
First trimester 0.32-1.81 Second trimester -14.88 -32.96 to 3.20 -1.47 -22.16 to 19.23
0.31-1.85 Third trimester 11.52 -7.59 to 28.18 11.52 -9.27 to 32.31
0.34-1.76 -9.05 -27.53 to 9.43 -12.56 -33.84 to 8.73
* Adjusted for infant sex, gestation week, maternal age, married status, maternal race,
maternal education, cigarette smoking during pregnancy, conception season, prenatal care,
and weight gained during pregnancy.
31


CHAPTER 5
DISCUSSION
In this study, LBW was associated with maternal exposure to high CO during the
third trimester of pregnancy. This result is generally consistent with several previous
studies, which include the studies conducted in Los Angeles, and six northeastern
cities of the United States. These studies reported significant associations between
CO exposure and LBW during the third trimester (Rite and Yu, 1999; Wilhelm and
Ritz, 2005), or the second to third trimester (Maisonet et al., 2001). The results in
Wilhelm and Ritzs study of 1994-2000 births generally confirm their previous
observations for the period of 1989-1993 (Rite and Yu, 1999). They observed the
stronger associations between CO and LBW in the third trimester of pregnancy when
restricting their analyses to women who resided close to CO air monitoring stations.
However, these two studies were unable to adjust maternal weight gain during
pregnancy and maternal smoking. This study was able to control these risk factors
that were not adjusted in the Los Angeles studies, and still found an association
between high CO exposure and LBW in the third trimester of pregnancy. Another two
studies conducted in Korea and Brazil found the link between CO exposure and LBW
in the first trimester (Ha et al., 2001; Gouveia et al., 2004). Although the effects
period is still undergoing debate, Wilhelm and Ritz (2005) suggested that the entire
pregnancy period may influence term LBW for CO exposure. Also, they pointed out
that the accumulation of exposure throughout pregnancy may affect fetal growth
possibly in addition to peak exposures during especially vulnerable periods (Wilhelm
and Ritz, 2005).
However, the results of this study are inconsistent with the results of the studies
conducted in Denver. California, northern Nevada, and Vancouver. The Denver study
(Alderman et al, 1987) found no clear association between ambient CO exposure in
32


the third trimester of pregnancy and the risk of LBW. In the early 1970s, a study
conducted in Los Angeles reported a lower mean birth weight for babies whose
mothers lived in areas of high ambient CO air pollution, which routinely exceeded
300 ppm (Air Pollution Control District, County of Los Angeles, 1972). Compared to
the study in Los Angeles during the 1970s, the CO levels reported for Denver during
the study period, whose median ranged from 0.5 to 3.6 ppm, were considerable lower.
Nevertheless, Alderman et al. (1987) found a small but insignificant increase in the
risk of LBW for average CO exposure greater than 3 ppm (OR=1.5; 95%CI, 0.7-3.5),
and they suggested that low levels of ambient CO may have an effect on birth weight.
However, this effect could not be detected in the Denver study because of the
relatively small study size.
This study had several strengths. First, the study accounted for most of the risk
factors considered previously as confounders in the Los Angeles (maternal smoking)
and San Paulo (maternal smoking and weight gained during pregnancy) studies. Maternal
smoking contributes significantly to the risk of LBW because CO binds with
hemoglobin to produce elevated levels of HbCO in smokers, decreasing the oxygen
available to the fetus (Davies ct al, 1979). The advantage of the ability to control
maternal smoking in the study of ambient CO exposure and LBW is significant, and
this can eliminate the effect maternal smoking may contribute to the outcome of the
analysis of exposure to ambient CO pollution. Second, this study provided a more
accurate exposure assessment for individual mothers by using the IDW method to
predict CO concentrations and examined the mothers who lived within a 2 mile radius
of the stations, providing much more homogeneous results with respect to social,
economic, and behavioral risk factors (Rite and Yu, 1999). Although the analyses
controlled for several potential confounders, information adjusting for some
additional known increasers of the risk of low birth weight, such as maternal nutrition,
social economic status (SES), and occupational exposures were not available. The
results with respect to social and economic risk factors may be unlikely to attribute to
33


variations in those risk factors in my study areas.
The primary weakness of this study was due to the limited number of air
monitoring stations and the limited data for other air pollutants in the study area. Thus,
the study was unable to examine the interaction between CO and other ambient air
pollutants. However, Ritz and Yu (1999) provided evidence that ambient CO is the
most consistent and important predictor of LBW as shown in the study conducted in
Los Angeles. Also, they found the same association between CO exposure and LBW
in the multi-pollutant model.
Potential misclassification of exposure is the most important source of error in
this study. First, although GIS was used to estimate the CO concentration within a 2
mile radius of the stations, the locations of pollutant sources and the atmospheric
influence upon the dispersion and ultimate concentrations of CO were ignored (Roger
et al, 1999). In addition, ambient CO pollution associated with traffic was also not
considered. The application of certain modeling techniques which consider
atmospheric influence and traffic related air pollution may improve this
misclassification.
Second, due to the use of simulation data rather than on-sitc measurement data to
determine the maternal exposure to the ambient CO air pollutant, the inability to
control indoor air pollution and to perform individual level exposure assessments are
another limitations in the study. Gas- and wood-burning stoves and second-hand
cigarette smoking are the most important indoor sources of CO (Ott et al, 1992).
However, if no indoor sources exist, outdoor levels of CO determine a large
percentage of indoor CO levels (Weinberg and Wilcox, 1998). In Denver, the annual
average temperature is about 50 F, and the average temperature is about 33 F during
the winter time (CityRating). Nevertheless, residential air exchange rates may be
lower in Denver areas less than in other warmer cities because people in Denver may
not open their windows and doors as often. Thus, indoor CO pollution and passive
cigarette smoking might influence the results of this study.
34


Third, in this study, it was assumed that mothers did not move during pregnancy
and they spent most of their time near monitoring stations. However, the mothers in
this study cohort may spend substantial amounts of timing outside the monitoring
districts during pregnancy. This is the common limitation for previous studies
because this information is not contained in birth registry databases.
Finally, in this ecological study, the residential addresses were not available at the
point location. Thus, the results of this study associated with block level may not
apply to individuals. Additionally, the monthly period may not correlate directly to
gestation period. For example, if a woman conceived on a specific day of the month,
the concentration of CO may be off set by how many days into the month conception
occurred. However, the changing in level of CO from month to month is minimal.
35


CHAPTER 6
CONCLUSIONS
The results suggest that there was an adverse effect on high CO exposure and
birth weight in the third trimesters of pregnancy during this time period. This finding
is similar to the study conducted by Wilhelm and Ritz, 2005 in Los Angles.
Compared with the other previous studies, this Study improved the estimation of CO
concentration by performing IDW interpolation method and was also able to control
for other important risk factors, such as maternal smoking during pregnancy, weight
gain, and prenatal care. Relatively speaking, Denver has somewhat low
concentrations of CO compared with a city like Los Angles. However, in this study
an association between high CO exposure and birth weight persisted even at the lower
levels. Future studies may improve the exposure assessment methods to have better
estimate personal levels of exposure from air monitoring data or collecting individual
exposure information. Study of epigenetics may also provide more biological
evidence on CO exposure and LBW. Obtaining detailed information on confounders
of CO exposure and LBW would definitely reduce the uncertainty of the study
outcomes.
36


BIBLIOGRAPHY
Air Pollution Control District, County of Los Angeles. (1972) 1972 Annual Report of
Meteorology, Air Pollution Effects, and Contaminant Maxima at Los Angeles
County. Los Angeles. California: Air Pollution Control District.
Alderman, B. W., Baron, A. E., & Savitz, D. A. (1987). Maternal exposure to
neighborhood carbon monoxide and risk of low infant birth weight. Public Health
Reports, 106, 410-414.
Astrup, P. Olsen, H. M., Trolle, D., & Kjeldsen, K. (1972). Effect of moderate carbon
monoxide exposure on fetal development. Lancet, 2, 1220-1222.
Bchcra, D., Dash, S., & Malik, S. K. (1988). Blood carboxyhacmoglobin levels
following acute exposure to smoke of biomass fuel. The Indian Journal of Medical
Research, 88, 522-524.
Behrman, R. F., Fisher, D. E., & Patton, J. (1971). Air pollution in nurseries:
Correlation with decrease in oxygen-carrying capacity of hemoglobin. Journal of
Pediatric, 78, 1050-1054.
Bobak, M. (2000). Outdoor air pollution, low birth weight, and prematurely.
Environmental Health Perspectives, 108,173-176.
Bobak, M., & Leon, D. A. (1999). Pregnancy outcomes and outdoor air pollution: an
ecological study in districts of the Czech Republic 1986-8. Occupational and
Environmental Medicine, 56, 539-543.
Boy, E., Bruce, N., & Delgado, H. (2002) Birth weight and exposure to kitchen wood
smoke during pregnancy in rural Guatemala. Environmental Health Perspective, 110,
109-114.
Boy, E., Bruce, N., & Delgado, H. (2002). Birth weight and exposure to kitchen wood
smoke during pregnancy in rural Guatemala. Environmental Health Perspectives,
110, 109-114.
Brenner, H., Savitz, D. A., Jockel, K., & Greenland, S. (1992). Effects of non
differential exposure misclassification in ecologic studies. American Journal of
Epidemiology, 135, 85-95.
Bruce, N., Perez-Padilla, R., & Albalak, R. (2000). Indoor air pollution in developing
37


countries: A major environmental and public health challenge. Bulletin of the World
Health Organization, 78, 1080-1092.
Butler, N. R., & Alberman, E. D. (1969). Perinatal Problems. The second report of
the 1958 British perinatal mortality survey under the auspices of the national
birthday fund. Edinburgh, Livingstone, Ltd.
Butler, N. R., Goldstein, H., & Ross, E. M. (1972). Cigarette smoking in pregnancy:
Its influence on birth weight and perinatal mortality. British Medical Journal, 2,
127-130.
Chen, L., Yang, W., Jennison, L. B., Goodrich, A., & Omaye, T. S. (2002). Air
pollution and birth weight in northern Nevada, 1991-1999. Inhalation Toxicology,
14, 141-157.
City Rating Agency. (2002). Average Temperature: Denver. Retrieved April 2,2007,
from the City Rating Agency
website :http://www.citvratingcom/citvtemperature.asp?Citv=Denver
Cnattingius, S. (1989). Does age potentiate the smoking-related risk of fetal growth
retardation? Early Human Development, 20, 203-211.
Cnattingius, S., Axeisson, O., Eklund, G., & Lindmark, G. (1985). Smoking, maternal
age, and fetal growth. American Journal of Obstetrics Gynecology, 66, 449-452.
Cnattingius, S., Forman, M. R., Berendes, H. W., Graubard, B. I., & Isotalo, L.
(1993). Effect of age, parity, and smoking on pregnancy outcome: a population-
based study. American Journal of Obstetrics Gynecology, 168, 16-21.
Comstock, G. W., & Lundin, F. E. (1967). Parental smoking and perinatal mortality.
American Journal of Obstetrics and Gynecology, 98, 708-718.
David, W. W., Lester, Y., & Susan, A. P. (2004). Comparison of spatial interpolation
methods for the estimation of air quality data. Journal of Exposure Analysis and
Environmental Epidemiology, 14, 404-415.
Davies, J. M Latto, I. P Jones, J. G Veale, A., & Wardrop, C. A. (1979). Effects of
stopping smoking for 48 hours on oxygen availability from the blood: a study on
pregnant women. British Medical Journal, 2, 355-356.
De Koning, H. W., Smith, K. R., & Last, J. M. (1985). Biomass fuel combustion and
38


health. Bulletin of the World Health Organization, 63, 11-26.
Dejmek, J., Selevan, S. G., Benesm, I., Solansky, I., & Sram, R. J. (1999). Fetal
growth and maternal exposure to particulate matter during prcgnanct. Environmental
Health Perspectives, 107, 475-480.
Dejmek, J., Solansky, I., Benes, I., Lenicek, J., & Sram, R. J. (2000). The impact of
polycyclic aromatic hydrocarbons and fine particles on pregnancy outcome.
Environmental Health Perspective, 12, 1159-1162.
Department of Public Health & Environment. (2000) Weighing in on solutions to Low
Birth Weight problem in Colorado.
Dugandzic, R., Dodds, L., Stieb, D., & Marc, S. D. (2006). The association between
low level exposures to ambient air pollution and term low birth weight: a
retrospective cohort study. Environmental Health: a global access sciences source,
5, 1-8.
Everson, R. B., Randerath, E., Santella, R. M., Avitts, T. A., Weinstein, I. B., &
Randerath, K. (1988). Quantitative associations between DNA damage in human
placenta and maternal smoking and birth weight. Journal of the National Cancer
Institute, 80, 567-576.
Fabia, J. (1973). Cigarettes pendant la grossess, poids de naissance et mortalite'
perinatale. Canadian Medical Association, 109, 1104-1107.
Fox, S. H., Koepsell, T. D., & Daling, J. R. (1994). Birth weight and smoking during
pregnancy-effect modification by maternal age. American Journal of Epidemiology,
739,1008-1015.
Frazier, T. M., Davis, G. H., Goldstein, H., & Goldberg, I. D. (1961). Cigarette
smoking and prematurity: A prospective study. American Journal of Obstetrics and
Gynecology, 81, 988-996.
Garvey, D., & Longo, L. D. (1978). Chronic low level maternal carbon monoxide
exposure and feta! growth and development. Biohg}> of Reproduction, 19, 8-14.
Gouveia, N., Bremner, S. A., & Novaes, H. M. (2004). Association between ambient
air pollution and birth weight in Soa Paulo, Brazil. Journal of Epidemiology
Community Health, 58, 11-17.
39


Gritz, E. R. (1980). Alcohol and drug problems in women. In Research Advances in
Alcohol and Drug problems. New York: Plenum Press.
Ha, E. H., Hong, Y. C., Lee, B. E., Woo, B. H., Schwartz, J., & Christian!, D. C.
(2001). Is air pollution a risk factor for low birth weight in Seoul? Epidemiology, 12,
643-48.
Hamby, D. M. (2002). The Gaussian atmospheric transport model and its sensitivity
to the joint frequency distribution and parametric variability. Health physics, 82, 64-
73.
Hanna, S. R., & Briggs, G.A, (1982). National Technical Information Service
Handbook on Atmospheric Diffusion. Springfield. VA: US Department of
Commerce.
Holland, J. Z. (1953). A meterological survey of the oak ridge area: final report
covering the period 1948-1952. Oak Ridge, TN: Weather Bureau. USAEC report
ORO-99.
Jerrett, M., Arain, A., Kanaroglou, P., Beckerman, B., Potoglou, D., Sahsuvaroglu, T.,
Morrison, J., & Giovis, C. (2005). A review and evaluation of intraurban air
pollution exposure models. Journal of Exposure Analysis and Environmental
Epidemiology, 75,185-204.
Joseph, K. S., & Kramer, M. S. (1996). Review of the evidence on fetal and early
childhood antecedents of adult chronic disease. Epidemiologic Reviews, 18, 158
174.
Kannan, S., Misra, D. P., Dvonch, J. T., & Krishnakumar, A. (2006). Exposures to
airborne particulate matter and adverse perinatal outcomes: a biologically plausible
mechanistic framework for exploring potential effect modification by nutrition.
Environmental Health Perspectives, 114, 1636-1642.
Kramer, M. S. (1987). Determinants of low birth weight: methodological assessment
and meta-analysis. Bull World Health Organization, 65, 663-737 a.
Kramer, M. S. (1987). Intrauterine growth and gestational duration determinants.
Pediatrics, 80, 502-511 b.
Kramer, M. S., Olivier, M., McLean, F. H., Dougherty, G. E., Willis, D. M., & Usher,
R. H. (1990). Determinants of fetal growth and body proportionality. Pediatrics, 86,
40


18-26.
Landgren, O. (1996). Environmental pollution and delivery outcome in southern
Sweden: a study with central registries. Ada Paediatrica, 85, 1361-1364.
Lee, B. E., Ha, E. H., Park, H. S., Kim, Y. J., Hong, Y. C., Kim, H., & Lee, J. T.
(2003). Exposure to air pollution during different gestational phase contributes to
risks of low birth weight. Human Reprodudive, 18, 638-643.
Leem, J. H., Kaplan, B. M., Shim, Y. K., Pohl, H. R., Gotway, C. A., Bullard, S. M.,
Rogers, J. F., Smith, M. M., & Tylenda, C. A. (2006). Exposure to air pollutants
during pregnancy and preterm delivery. Environmental Health Perspedives, 114,
905-910.
Longo, L. D. (1977). The biological effects of carbon monoxide on the pregnant
woman, fetus, and newborn infant. American Journal of Obstetrics Gynecology, 129,
69-103.
Longo, L. D. (1977). The biological effects of carbon monoxide on the pregnant
woman, fetus, and newborn infant. American Journal of Obstetrics and Gynecology,
129, 69-103.
Maisonet, M., Bush, T. J., Correa, A., & Jaakkola, J. K. (2001). Relation between
ambient air pollution and low birth weight in the northeastern United States.
Environmental Health Perspectives, 3, 351-356.
Mavalankar, D. V., Trivedi, C. R., & Gray, R. H. (1991) Levels and risk factors for
perinatal mortality in Ahmedabad, India. Bulletin of the World Health Organization,
69, 435-442.
Meade, P. J., & Pasquill, F. (1951). A study of the average distribution of pollution
around Staythorpe. Intrnational Journal of Air Pollution, 11, 60-70.
Miller, H. C., & Hassanein, K. (1974). Maternal smoking and fetal growth of full
term infants. Pediatric Research, 8, 960-963.
Mishra, V., Dai, X., Smith, K. R., & Mika, L. (2004). Maternal exposure to biomass
smoke and reduced birth weight in Zimbabwe. Annals of Epidemiology, 14, 740-747.
Mongelli, M. & Biswas, A. (2001). A fetal growth standard derived from multiple
modalities. Early Human Development, 60, 171-177.
41


Morgenstem, H., & Thomas, D. (1993). Principles of study design in environmental
epidemiology. Environmental Health Perspectives, 4, 23-38.
National Academy of Sciences, National Research Council. (1977). Committee on
Medical and Biological Effects of Environmental Pollutants: Carbon monoxide.
Washington: National Academy Press.
Niermeyer, S., Williford, D., Asmus, I., Honigman, B., Moore, L. G., Wells, C., &
Egbert, M. (2006). Low Birth Weight in Colorado: analysis using geographic
information systems. 2006 ESRI Health GIS Conference Poster.
Ott, W. R., Mage, D. T., & Thomas, J. (1992). Comparison of microenvironmental
CO concentrations in two cities for human exposure modeling. Journal of Exposure
Annual Environmental Epidemiology, 2, 249-267.
Parker, J. D., Woodruff, T. J., Basu, R., & Schoendorf, K. C. (2005). Air pollution
and birth weight among term infants in California. Pediatrics, 115,121-28.
Pereira, L. A., Loomis, D., Conceicao, G. M., Braga, A. L., Areas, R. M., Kishi, H. S.,
Singer, J. M., Bohm, G. M., & Saldiva, P. H. (1998). Association between air
pollution and intrauterine mortality in Sao Paulo, Brazil. Environmental Health
Perspectives, 106, 325-329.
Perera, F. P., Rauh, V., Tsai, W. Y., Kinney, P., Camann, D., Barr, D., Bemert, T.,
Garfmkel, R Tu, Y. H Diaz, D Dietrich, J., & Whyatt, R. M. (2003). Effects of
transplacental exposure to environmental pollutants on birth outcomes in a
multiethnic population. Environmental Health Perspectives, 111, 201-205.
Ritz, B. & YU, F. (1999). The effect of ambient carbon monoxide on low birth weight
among children bom in southern California between 1989 and 1993. Environmental
Health Perspectives, 107, 17-25.
Ritz, B., Yu, F., Chapa, G., & Fruin, S. (2000). Effect of air pollution on preterm birth
among children bom in southern California between 1989 and 1993. Epidemiology,
11, 502-511.
Ritz, B., YU, F., Fruin, S., Chapa, G., Shaw, G. M., & Harris, J. A. (2002). Ambient
air pollution and risk of birth defects in Southern California. American Journal of
Epidemiology, 155, 17-25.
Roger, J. F., & Dunlop, A. L, (2006). Air Pollution and very Low Birth Weight Infants:
42


A Target Population? Pediatrics, 118, 156-161.
Rogers, J. F., Killough, G. G., Thompson, S. J., Addy, C. L., Mckeown, R. E., &
Cowcn, D. J. (1999). Estimation environmental exposure to sulfur dioxide from
multiple industrial sources for a case-control study. Journal of Exposure Analysis
and Environmental Epidemiology, 9, 535-545.
Silbergeld, E., & Tonat, K. (1994). Investing in prevention: opportunities to prevent
disease and reduce health care costs by identifying environmental and occupational
causes of noncancer disease. Toxicology and Industrial Health, 10, 675- 827.
Simpson, W. J. (1957). A preliminary report on cigarette smoking and the incidence
of prematurity. American Journal of Obstetrics and Gynecology, 73, 807-815.
Sram, R. J., Binkova, B., Rassncr, P. Topinka, J., & Dcjmck, J. (1999). Adverse
reproductive outcomes from exposure to environmental mutagens. Mutation
Research, 428, 203-215.
Topinka, J., Binkova, B., Mrackova, G., Stavkova, Z., Peterka, V., Benes, I., Dejmek,
J., Lenicek, J., Pilcik, T., & Sram, R. J. (1997). Influence of GSTM1 and NAT2
genotypes on placental DNA adducts in an environmentally exposed population.
Environmental Molecular Mutagen, 30, 184-195.
U.S. Environmental Protection Agency. (1997). Environmental Criteria and
Assessment Office: Air quality criteria for carbon monoxide. Publication No. EPA-
600/8-79-022, Research Triangle Park, NC.
Wang, X., Ding, H., Ryan, L., & Xu, X. (1997). Association between air pollution
and low birth weight: a community-based study. Environmental health perspectives,
105, 514-520.
Weinber, C. R., & Wilcox, A. J. (1998). Reproductive epidemiology. In: Modern
Epidemiology. 2nd Ed. Philadelphia: Lippincott-Raven Publishers.
Wells, L. L. (1933). The prenatal effect of carbon monoxide on albino rats and the
resulting neuropathology. Biologist, 15, 80-81.
Wen, S. W., Golsenberg, R. L., Cutter, G. R., Hoffman, H. J., Cliver, S. P., Davis, R.
O., & DuBard, M. B. (1990). Smoking, maternal age, fetal growth, and gestational
age at delivery. American Journal of Obstetrics Gynecology, 162, 53-58.
43


Wilhelm, M., & Ritz, B. (2003). Residential proximity to traffic and adverse birth
outcomes in Los Angeles county, California, 1994-1996. Environmental Health
Perspectives, 111, 207-216.
Wilhelm, M., & Ritz, B. (2005). Local variations in CO and particulate air pollution
and adverse birth outcomes in Los Angeles county, California, USA. Environmental
Health Perspectives, 13, 1212-1221.
Williams, L., Spence, A. M., & Tideman, S. C. (1977). Implication of the observed
effect of air pollution on birth weight. Social Biology, 24, 1-9.
World Resources Institute, UNEP, UNDP, World Bank. (1998). 1998-99 world
resources: a guide to the global environment. Oxford: Oxford University Press.
Woutcrs, E. J., Jong, P. A., Comclisscn, P. J., Kurvcr, P. H., Ocl, W. C., & Wocnscl,
C. L. (1987). Smoking and low birth weight: absence of influence by carbon
monoxide? European Journal of Obstetrics, Gynecology, and Reproductive Biology,
25, 35-41.
Xu, X., Ding, H., & Wang, X. (1995). Acute effects of total suspended particles and
sulfur dioxides on preterm delivery: a community based cohort study. Archives of
Environmental Health, 50, 407-415.
44