University of Southern California Dissertations and Theses

Personal exposure to particulate matter PM2.5 sources during pregnancy and birthweight

(USC Thesis Other)

Personal exposure to particulate matter PM2.5 sources during pregnancy and birthweight

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content PERSONAL EXPOSURE TO PARTICULATE MATTER PM 2.5 SOURCES DURING
PREGNANCY AND BIRTHWEIGHT
By
Karl O’Sharkey

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements of the Degree
DOCTOR OF PHILOSOPHY
(EPIDEMIOLOGY)

August 2022

ii
DEDICATION

I dedicate this work to my wife, Nathalie. All my achievements in life are because of you.

I also dedicate this dissertation to my dog, Murray. Rest in peace, my little mate.

iii
TABLE OF CONTENTS
DEDICATION ............................................................................................................................................... ii
LIST OF TABLES ......................................................................................................................................... v
LIST OF FIGURES ...................................................................................................................................... vi
ABSTRACT ................................................................................................................................................ vii
OVERVIEW .................................................................................................................................................. 1
REFERENCES .......................................................................................................................................... 7
CHAPTER 1
BACKGROUND AND REVIEW ............................................................................................................... 10
MADRES STUDY .................................................................................................................................. 10
HEALTH DISPARITIES ........................................................................................................................ 11
THE IMPACT OF BIRTHWEIGHT ON HEALTH .............................................................................. 12
BIOMECHANISM OF LOW BIRTHWEIGHT IMPACTS ON HEALTH .......................................... 14
OUTDOOR PARTICULATE MATTER AIR POLLUTION ................................................................ 17
PERSONAL PM 2.5 EXPOSURE ............................................................................................................. 19
ASSOCIATION BETWEEN PRENATAL EXPOSURE TO PM 2.5 AND BIRTHWEIGHT ................ 20
ASSOCIATION BETWEEN OUTDOOR PM 2.5 SOURCES AND COMPONENTS AND
BIRTHWEIGHT ..................................................................................................................................... 22
EXPOSURE MEASUREMENT ERROR OF SECONDHAND SMOKE IN PREGNANCY STUDIES
................................................................................................................................................................. 24
SUMMARY AND OVERALL AIMS .................................................................................................... 26
REFERENCES ........................................................................................................................................ 28
CHAPTER 2
IN-UTERO PERSONAL EXPOSURE TO PM 2.5 IMPACTED BY INDOOR AND OUTDOOR
SOURCES IN THE MADRES COHORT .................................................................................................. 38
ABSTRACT ............................................................................................................................................ 38
INTRODUCTION ................................................................................................................................... 39
MATERIALS AND METHODS ............................................................................................................ 41
Study Population ................................................................................................................................................. 41
Personal PM2.5 Exposure Monitoring ................................................................................................................. 42
Birthweight ......................................................................................................................................................... 43
Questionnaire and Other Covariate Data ............................................................................................................ 43
Statistical Analysis ............................................................................................................................................. 45
RESULTS ............................................................................................................................................... 48
DISCUSSION ......................................................................................................................................... 57
REFERENCES ........................................................................................................................................ 62
SUPPLEMENTARY MATERIALS ....................................................................................................... 68

iv
CHAPTER 3
EFFECTS OF IN-UTERO PERSONAL EXPOSURE TO PM 2.5 SOURCES ANDCOMPONENTS ON
BIRTHWEIGHT ......................................................................................................................................... 71
ABSTRACT ............................................................................................................................................ 71
INTRODUCTION ................................................................................................................................... 73
MATERIALS AND METHODS ............................................................................................................ 75
Study Population ................................................................................................................................................. 75
Personal PM2.5 Exposure Monitoring ................................................................................................................. 76
Personal PM2.5 Sources ....................................................................................................................................... 77
Birthweight Outcome ......................................................................................................................................... 79
Covariate Data .................................................................................................................................................... 79
Statistical Analysis ............................................................................................................................................. 81
RESULTS ............................................................................................................................................... 82
DISCUSSION ......................................................................................................................................... 89
REFERENCES ........................................................................................................................................ 95
SUPPLEMENTARY MATERIALS ..................................................................................................... 100
CHAPTER 4
A COMPARISON OF MEASURED AIRBORNE AND SELF-REPORTED SECONDHAND SMOKE
EXPOSURE IN THE MADRES PREGNANCY COHORT STUDY ...................................................... 102
ABSTRACT .......................................................................................................................................... 102
INTRODUCTION ................................................................................................................................. 103
METHODS AND MATERIALS .......................................................................................................... 107
Study Population ............................................................................................................................................... 107
Personal PM2.5 and Environmental Tobacco Smoke (ETS) Exposure Monitoring .......................................... 108
Self-Report Secondhand Smoke Exposure Questionnaire and EMR Variables .............................................. 110
Statistical Analysis ........................................................................................................................................... 111
RESULTS ............................................................................................................................................. 112
DISCUSSION ....................................................................................................................................... 117
REFERENCES ...................................................................................................................................... 123
CHAPTER 5
SUMMARY AND FUTURE RESEARCH DIRECTIONS ...................................................................... 129
Conclusions and Implications ............................................................................................................... 132
Future Studies and Research Directions ................................................................................................ 134
REFERENCES ...................................................................................................................................... 137

v

LIST OF TABLES
Table 2.1. Descriptive Statistics of Study Participants (N=205). ............................................................... 48

Table 2.2. Regression Results for Base Model of PM 2.5 and Birthweight (N = 205). ................................ 51

Table 2.3. Estimated Change in Birthweight (g) per 1 SD Increase in Personal PM 2.5 from Interaction
Analyses (N = 204). ............................................................................................................... 53

Table S.2.1. Bivariate Analysis of Personal PM 2.5 and Birthweight by Categorical Participant
Characteristics. (N = 205). ..................................................................................................... 68

Table S.2.2. Bivariate Analysis of Personal PM 2.5 and Birthweight by Participant Characteristics (N =
205). ........................................................................................................................................ 70

Table 3.1. Sample Participant Characteristics (N = 201). ........................................................................... 83

Table 3.2. Summary Statistics of PM 2.5 Sources and Components Concentrations (N = 201). .................. 84

Table 3.3. Spearman’s Correlation Coefficients for Major Contributing Sources of Personal PM 2.5 (N =
201). ........................................................................................................................................ 86

Table 3.4. Single- and Two-pollutant Associations Between PM 2.5 Sources and Birthweight. .................. 87

Table 3.5. Estimated Change in Birthweight (g) per 1 SD Increase in Pollutant by Infant Sex (N = 198).
................................................................................................................................................ 89

Table S.3.1. Personal PM 2.5 Sources by Key Sample Demographics (N = 201). ..................................... 100

Table S.3.2. Spearman’s Correlation Coefficients for High-Loading Components of the Six Personal
PM 2.5 Sources (N = 201). ...................................................................................................... 101

Table S.3.3. Effect estimates of major personal PM 2.5 Sources on Birthweight in Full Term Births Only.
.............................................................................................................................................. 101

Table 4.1. Summary of Study Participants by Measured Personal PM 2.5 and ETS Concentrations……..113

Table 4.2. Summary of Environment/Household Characteristics by Measured Personal PM 2.5 and ETS
Concentrations. ..................................................................................................................... 114

Table 4.3. Summary of Measured Personal PM 2.5 and ETS Concentrations by Self-Reported SHS
Exposure Questionnaire Responses. .................................................................................... 116

Table 4.4. Bivariate Analysis of ETS by Personal and Ambient PM 2.5. ................................................... 117

vi

LIST OF FIGURES

Figure 1.1. Maternal, placental, and fetal factors in fetal growth impacted by environmental exposures. 15

Figure 2.1. Relationship of Personal PM 2.5 and Outdoor PM 2.5 in (a) the 48-hour Monitoring Period and
(b) the Third Trimester of Pregnancy. .................................................................................... 51

Figure 2.2. Predicted Relationship of Personal PM 2.5 Exposure on Birthweight (a) Overall, (b) by Type of
Home, (c) Time Spent Indoors, and (d) Air Conditioner Use at Home, for each level of the
interaction variable where applicable. .................................................................................... 56

Figure 3.1. Associations between personal PM 2.5 components and birthweight, controlling for personal
PM 2.5 mass. ............................................................................................................................. 88

vii

ABSTRACT
Low birthweight (LBW) is an important birth health metric, with studies finding associations of
LBW with several adverse health outcomes in both early and later life, including metabolic diseases,
cardiovascular disease, cognitive impairment, and infant mortality. Previous studies generally support a
weak to moderate association between outdoor air pollution and specifically particulate matter with
aerodynamic diameter less than 2.5µm (PM 2.5) and decreased birthweight. However, previous results are
mixed, likely due to measurement error introduced by estimating personal exposure from ambient data, and
from differences in the major contributing sources of PM 2.5, which likely have different toxicities.
Additionally, while outdoor PM 2.5 is an important research question due to this criteria air pollutant being
regulated and shown to have important population-wide adverse effects on health, personal PM 2.5 exposure
consists of outdoor and indoor PM 2.5 sources as a result of individuals’ time-activity patterns (i.e., time
spent indoors or outdoors), various behaviors, and infiltration of outdoor PM 2.5 indoors. This total external
exposure in the personal breathing zone is best captured by personal monitoring. Very few studies have
been able to investigate the effects of personal PM 2.5 exposure during pregnancy on birthweight and whether
that differs by the major sources contributing to it, particularly in the 3
rd
trimester where most fetal weight
gain occurs. This dissertation aimed at addressing the following three research questions to better
understand the effect of in-utero exposure to personal PM 2.5 and its major contributing sources on
birthweight in the 3
rd
trimester within a health disparities population in Los Angeles, CA:
1) Does the effect of personal PM 2.5 on birthweight vary based on whether it was mostly impacted by
PM 2.5 sources of indoor vs outdoor origin?
2) What is the effect of specific sources of personal PM 2.5 on birthweight, as characterized by their
chemical composition?
3) Is personal exposure to secondhand smoke (SHS), a major source of personal PM 2.5 and a known
risk factor for reductions in birthweight, adequately assessed through self-reported SHS
questionnaires?

viii
Within this dissertation, I present evidence that while total personal PM 2.5 exposure was not
significantly associated with birthweight, indoor versus outdoor origin, and more specific sources of
personal PM 2.5 appear to be adversely associated with birthweight. Home characteristics, including, more
time with windows open in the home, not using air conditioning (AC), candle/incense smoke exposure, and
more time spent outdoors, were associated with a more negative effect of personal PM 2.5 on birthweight.
This suggests that sources originating indoors versus outdoors and potentially different pathways may
impact this relationship, but it also highlights possible behavioral interventions to reduce risk. Additionally,
using a chemical speciation and factor analysis approach, sources of personal PM 2.5, such as fresh sea salt,
aged sea salt, and to a lesser degree, SHS and crustal sources, were associated with risk of decreased
birthweight. Furthermore, high-loading components of these personal PM 2.5 sources or factors and
exposures highly correlated with them, may drive these associations, with magnesium and sodium most
inversely associated with birthweight. Finally, I provide evidence that SHS, a major contributing source of
personal PM 2.5, may not be accurately measured by self-reported questionnaires which are typically used in
pregnancy studies, supporting a call for standardized SHS exposure questionnaires to increase accuracy in
exposure assessment and allow for harmonization and pooling of data and collaboration across cohorts.
Overall, this dissertation highlights a need for more health research and targeted regulation and/or
guidelines using PM 2.5 sources and components considering PM 2.5 is a complex mixture of organic and
inorganic particles, which seem to differ in toxicity.

1

OVERVIEW
Air pollution, including particulate matter with aerodynamic diameter < 2.5µm (PM 2.5), is a mixture
of solid and gas particles, emitted or formed from both man-made and natural sources that can be hazardous
to human health. PM 2.5 exposure leads to an estimated 4.2 million and 3.8 million deaths every year from
ambient and indoor (household) air pollution, respectively (Prüss-Üstün et al., 2016). These negative health
outcomes are often disparate by socio-demographics, including, race/ethnicity, income, and education (Bell
& Ebisu, 2012). Given that health disparities often exist as early as infancy, their origins are hypothesized
to start in the earliest stages of life. The early origin of health and disease hypothesis has garnered a great
deal of attention over the years, due to the concept that adverse pregnancy and early-life exposures lead to
a host of non-communicable diseases through the rest of early-life and into adulthood (Moore, 2017; Barnes
& Ozanne, 2011). Decreases in birthweight, including a commonly studied birth outcome, low birthweight
(LBW; typically defined as birthweight <2,500 grams), is associated with infant mortality (Vilanova et al.,
2019; Watkins et al., 2016), type-2 diabetes (Jornayvaz et al., 2016), cognitive development (Upadhyay et
al., 2019; Whitaker et al., 2006), and cancer (Cook et al., 2010; Harder et al., 2010) later in life. While most
of the literature has concentrated on LBW, high birthweight (macrosomia; > 4,000 grams) is also linked to
obesity (Schellong et al., 2012), type-2 diabetes (Harder et al., 2007), and cancer (Caughey & Michels,
2009; Harder et al., 2010) risk later in life, highlighting that both low and high birthweight may impact
lifetime health.
Environmental exposures, such as PM 2.5, impact pregnancy and birth outcomes by impacting
oxidative stress, DNA methylation, and endocrine disruption (Clemente et al., 2016; Z. Li et al., 2019). The
literature generally supports a moderate negative association between outdoor PM 2.5 and birthweight and
an increased risk of being LBW, especially in the 3
rd
trimester where the majority of fetal weight gain
occurs (Dadvand et al., 2014; Rich et al., 2015; Savitz et al., 2014; Schembari et al., 2015). However, results
have been inconsistent (X. Li et al., 2017; Madsen et al., 2010; Sun et al., 2016), possibly due to most
studies estimating outdoor residential PM 2.5 exposure (where individuals spend most of their time) using

2
models that typically incorporate ambient monitoring, remote sensing, and geospatial data (Gray et al.,
2010; Harris et al., 2014). While these are increasingly capable of capturing spatial variability in outdoor
air pollution, they inherently suffer from exposure measurement error which attenuates power to detect
health effects (Carroll, 2005). This is because individuals spend the majority of their time indoors, and their
“true” personal (external) exposure to PM 2.5 of outdoor origin is a result of the infiltration efficiency of
outdoor sources of PM 2.5 indoors and time-activity patterns most accurately captured by personal
monitoring (Gray et al., 2010; Kioumourtzoglou et al., 2014). Additionally, a recent review article on
particulate matter (PM) components and health concluded that there is an important need for exposure
models that cover both outdoor and indoor environments to estimate total exposure, particularly for PM
components (Yang et al., 2018). For that reason, there is a need to capture the “true” personal exposure
individuals experience in their external breathing zone by using personal exposure monitoring, and
additionally, parse out the fragment that is of outdoor origin, which is the segment that policy is able to
target for reductions.
Additionally, PM 2.5 is itself a mixture of organic and inorganic components, including carbons,
metals, and crustal elements, with different chemical/elemental components being associated with lower
birthweight risk to differing degrees (Bell et al., 2012; Ebisu & Bell, 2012; Laurent et al., 2016). Prior
efforts to investigate PM 2.5 sources and/or components of outdoor origin likely also suffer from exposure
misclassification and attenuated effects due to outdoor spatial variability of components, which are not
adequately captured using ambient monitor data (Bell et al., 2011). High-loading components or those that
are considered markers or tracers of more specific sources of PM 2.5, including motor vehicles and second-
hand smoking, are also of interest since they may be driving risk themselves (versus correlating with a
specific mixture of PM 2.5 represented by the sources). Therefore, we investigated the effect of specific PM 2.5
sources and their high-loading components on birthweight to elucidate which sources or components are
more adversely related to lower birthweight risk. This work may also help with potentially identifying more
targeted intervention possibilities for policymakers.

3
One particularly important source of PM 2.5 exposure is secondhand smoke (SHS) which is often
endemic in society, and exposure to SHS during pregnancy is strongly associated with decreased
birthweight risk (Hawsawi et al., 2015; Pogodina et al., 2009). SHS exposure is also disparate with regard
to race/ethnicity, education, and income (Tsai, 2018). Unfortunately, SHS classification often suffers from
measurement error from using self-reported questionnaire data to quantify or recall exposure, with just
17.9% (~35% in pregnancy studies) validating their questions against objective measures, such as
biomarkers and air pollution exposure monitors (Pérez-Ríos et al., 2013). Researchers have found large
discrepancies between self-reported SHS and measured objective measures in pregnancy studies
(Argalasova et al., 2019; DeLorenze et al., 2002; O’Connor et al., 1995; Xiao et al., 2018), highlighting the
need to evaluate self-report exposure assessments for accuracy. Several approaches have been used,
including biomarkers and air pollution monitors (Jaakkola & Jaakkola, 1997), with less invasive and more
harmonizable and accurate measures and approaches still needed. Therefore, we investigated self-reported
SHS exposure questions from multiple questionnaires representing different wording and time-points and
capturing various aspects of exposure (duration, proximity, intensity) against measured airborne
environmental tobacco smoke (ETS; an analytically measured optical carbon fraction strongly correlating
with tobacco smoke).
In chapter 1 of this dissertation, I detail the prevalence of LBW and the impact of reduced
birthweight on early and later-life health outcomes, highlighting potential biological mechanisms that
underlie the adverse health effects over the course of life. Following this, I provide background on outdoor
and personal PM 2.5, including describing its sources and chemical composition, spatial distribution, and
common assessment methods, before describing the current literature on the effect of prenatal exposure to
outdoor and personal PM 2.5 and its sources and components on birthweight. I then discuss the inherent
limitations of previous literature, namely, the limited investigation into specific personal PM 2.5 components
and sources and the measurement error introduced from using ambient air pollution estimates to estimate
an individual’s exposure to PM 2.5 of outdoor origin. Next, I describe personal PM 2.5 exposure monitoring
and the strengths and limitations of using such an approach to improve the accuracy of external PM 2.5

4
exposure assessment in health studies. Finally, because of the importance of SHS as a source of personal
PM 2.5 exposure, and a known contributor to lower birthweight, I outlined current challenges and limitations
with SHS exposure assessment in epidemiological studies.
In chapter 2 (Study 1), I tested the hypothesis that total personal PM 2.5 is negatively associated
with birthweight. This was conducted using “gold standard” integrated personal exposure monitors over a
48-hour sampling period in the 3
rd
trimester, in a sample of Hispanic, lower-income pregnant mothers
within the MADRES pregnancy cohort at the University of Southern California. Multiple linear regression
was used to assess the effect of total personal PM 2.5 on birthweight, adjusting for key covariates.
Additionally, this study evaluated whether the effect of total personal PM 2.5 was modified by dominant or
major contributing sources of PM 2.5 (broadly defined as indoor vs outdoor in this analysis). Interaction
terms of total personal PM 2.5 and each indoor/outdoor source variable were added into the model and
evaluated as to whether the effect estimates differed, thereby suggesting potential differences in toxicity.
To our knowledge, this is one of the first studies to provide epidemiological evidence using personal PM 2.5
monitoring, in a health disparities population of primarily Hispanic, lower-income participants in Los
Angeles, CA. While total personal PM 2.5 exposure was not statistically significantly associated with
birthweight, there was evidence that multi-unit housing (vs. single-family homes), candle and/or incense
smoke, and greater outdoor source contributions to personal PM 2.5 were more strongly associated with
lower birthweight.
In chapter 3 (Study 2), I continued the theme of whether the different dominant or major
contributing sources or origins of PM 2.5 play a larger role in driving the negative effect of PM 2.5 exposure
during pregnancy on birthweight. However, this time I leveraged detailed PM speciation data to distinguish
sources more specifically based on their chemical fingerprint, as compared to the more general indoor vs
outdoor origin source categories in Study 1. The elemental and carbon components used in this analysis
were chosen based on them loading highly on the profiles of six main sources of personal PM 2.5 derived
with a Positive Matrix Factorization (PMF) source apportionment analysis (previously conducted by my
colleague Dr. Yan Xu). Multiple linear regression was used to estimate the effects of each source

5
contribution and high-loading component concentration individually, and within a multi-pollutant adjusted
approach, on birthweight adjusting for key covariates. Additionally, due to concerns that PM 2.5 mass may
be related to both the concentration of PM 2.5 components (for more abundant components) and birthweight,
analyses were conducted to adjust for PM 2.5 mass via two approaches. The first was by adding PM 2.5 mass
into the model as a covariate where possible (not collinear with other covariates), and the second by creating
component-specific residuals, thereby, holding PM 2.5 mass constant or adjusting for variation in the
components that is not driven by variation in PM 2.5 mass. This study found that personal PM 2.5 was not
significantly associated with decreased birthweight; however, the effect of major PM 2.5 sources on
birthweight was heterogeneous, with fresh sea salt and aged sea salt seemingly having the most negative
effect on birthweight, followed by crustal and SHS. Additionally, magnesium and sodium were associated
with the highest decrease in birthweight for sea salt sources, while measured airborne environmental
tobacco smoke (ETS) and brown carbon (BrC) were more negatively associated with birthweight than the
source itself. Adjusting for personal PM 2.5 mass did not materially alter the results. Finally, the association
of crustal and fuel oil sources with birthweight was modified by infant sex, with boys having a negative
association compared to a positive association for girls.
In chapter 4 (Study 3), I compared how well self-reported exposure to SHS from several recall-
based questionnaires asked in different time windows and using different wording to get at duration,
proximity, and intensity of exposure correlated with measurements of airborne ETS in PM 2.5 personal
samples collected in the 3
rd
trimester. I described the distribution of ETS concentrations amongst the
MADRES personal monitoring sub-study participants by key demographics and important
environment/household characteristics to elicit factors that may place individuals at greater risk of
exposure. Self-reported SHS exposure questions relating to the presence of smokers, intensity of smokers
(# of present smokers), and duration (hours of SHS exposure), were assessed via analysis of variance tests
to determine group mean differences in measured ETS concentrations. This study found no significant
association between measured airborne ETS and any self-report SHS exposure questions; however, asking
about the number of smokers nearby in the 48-hour monitoring period was most highly correlated with

6
measured ETS. This might suggest that recall of participants is more accurate when exposure is more
perceivable (such as by being close to a smoker or noticing smoke), and that less apparent exposure to SHS
through less obvious pathways (i.e., infiltration from neighboring residences) are more difficult to capture
via questionnaires. The larger implications of this work also shed light on how variable or poorly correlated
SHS exposure questions might be in what aspect of exposure they are truly capturing, posing important
challenges for data harmonization and limiting researchers’ ability to conduct larger, pooled analyses across
different contexts to more accurately understand the impacts of this ubiquitous exposure during pregnancy
on maternal and child health.
Finally, chapter 5 summarizes the main findings and implications of my dissertation, including
suggested future research directions. This dissertation found that the effect of personal PM 2.5 on birthweight
differed by major sources and components of PM 2.5, highlighting a need for more source-based health
research and regulation and guidelines aimed at reducing risk of decreased birthweight. Additionally, SHS
was a ubiquitous PM 2.5 source amongst study participants and self-reported questionnaires were not
sensitive to capturing this exposure. Therefore, there is a need for standardized SHS questionnaires that
have been validated and will provide more accurate exposure assessments but also allow data harmonization
across cohorts.

7
REFERENCES
Argalasova, L., Zitnanova, I., Vondrova, D., Dvorakova, M., Laubertova, L., Jurkovicova, J., Stofko, J.,
Weitzman, M., Waczulikova, I., & Simko, M. (2019). Self-Reported Exposure to ETS
(Environmental Tobacco Smoke), Urinary Cotinine, and Oxidative Stress Parameters in Pregnant
Women—The Pilot Study. International Journal of Environmental Research and Public Health,
16(9), 1656. https://doi.org/10.3390/ijerph16091656
Barnes, S. K., & Ozanne, S. E. (2011). Pathways linking the early environment to long-term health and
lifespan. Progress in Biophysics and Molecular Biology, 106(1), 323–336.
https://doi.org/10.1016/j.pbiomolbio.2010.12.005
Bell, M. L., Belanger, K., Ebisu, K., Gent, J. F., & Leaderer, B. P. (2012). Relationship Between Birth
Weight and Exposure to Airborne Fine Particulate Potassium and Titanium During Gestation.
Environmental Research, 117, 83–89. https://doi.org/10.1016/j.envres.2012.05.004
Bell, M. L., & Ebisu, K. (2012). Environmental Inequality in Exposures to Airborne Particulate Matter
Components in the United States. Environmental Health Perspectives, 120(12), 1699–1704.
https://doi.org/10.1289/ehp.1205201
Bell, M. L., Ebisu, K., & Peng, R. D. (2011). Community-level spatial heterogeneity of chemical
constituent levels of fine particulates and implications for epidemiological research. Journal of
Exposure Science & Environmental Epidemiology, 21(4), 372–384.
https://doi.org/10.1038/jes.2010.24
Carroll, R. J. (2005). Measurement Error in Epidemiologic Studies. In Encyclopedia of Biostatistics.
American Cancer Society. https://doi.org/10.1002/0470011815.b2a03082
Caughey, R. W., & Michels, K. B. (2009). Birth weight and childhood leukemia: A meta-analysis and
review of the current evidence. International Journal of Cancer, 124(11), 2658–2670.
https://doi.org/10.1002/ijc.24225
Clemente, D. B. P., Casas, M., Vilahur, N., Begiristain, H., Bustamante, M., Carsin, A.-E., Fernández, M.
F., Fierens, F., Gyselaers, W., Iñiguez, C., Janssen, B. G., Lefebvre, W., Llop, S., Olea, N.,
Pedersen, M., Pieters, N., Santa Marina, L., Souto, A., Tardón, A., … Nawrot, T. S. (2016).
Prenatal Ambient Air Pollution, Placental Mitochondrial DNA Content, and Birth Weight in the
INMA (Spain) and ENVIRONAGE (Belgium) Birth Cohorts. Environmental Health
Perspectives, 124(5), 659–665. https://doi.org/10.1289/ehp.1408981
Cook, M. B., Akre, O., Forman, D., Madigan, M. P., Richiardi, L., & McGlynn, K. A. (2010). A
systematic review and meta-analysis of perinatal variables in relation to the risk of testicular
cancer—Experiences of the son. International Journal of Epidemiology, 39(6), 1605–1618.
https://doi.org/10.1093/ije/dyq120
Dadvand, P., Ostro, B., Figueras, F., Foraster, M., Basagaña, X., Valentín, A., Martinez, D., Beelen, R.,
Cirach, M., Hoek, G., Jerrett, M., Brunekreef, B., & Nieuwenhuijsen, M. J. (2014). Residential
Proximity to Major Roads and Term Low Birth Weight: The Roles of Air Pollution, Heat, Noise,
and Road-Adjacent Trees. Epidemiology, 25(4), 518–525.
https://doi.org/10.1097/EDE.0000000000000107
DeLorenze, G. N., Kharrazi, M., Kaufman, F. L., Eskenazi, B., & Bernert, J. T. (2002). Exposure to
environmental tobacco smoke in pregnant women: The association between self-report and serum
cotinine. Environmental Research, 90(1), 21–32. https://doi.org/10.1006/enrs.2001.4380
Ebisu, K., & Bell, M. L. (2012). Airborne PM2.5 Chemical Components and Low Birth Weight in the
Northeastern and Mid-Atlantic Regions of the United States. Environmental Health Perspectives,
120(12), 1746–1752. https://doi.org/10.1289/ehp.1104763
Gray, S. C., Edwards, S. E., & Miranda, M. L. (2010). Assessing Exposure Metrics for PM and
Birthweight Models. Journal of Exposure Science & Environmental Epidemiology, 20(5), 469–
477. https://doi.org/10.1038/jes.2009.52

8
Harder, T., Plagemann, A., & Harder, A. (2010). Birth weight and risk of neuroblastoma: A meta-
analysis. International Journal of Epidemiology, 39(3), 746–756.
https://doi.org/10.1093/ije/dyq040
Harder, T., Rodekamp, E., Schellong, K., Dudenhausen, J. W., & Plagemann, A. (2007). Birth weight and
subsequent risk of type 2 diabetes: A meta-analysis. American Journal of Epidemiology, 165(8),
849–857. https://doi.org/10.1093/aje/kwk071
Harris, G., Thompson, W. D., Fitzgerald, E., & Wartenberg, D. (2014). The association of PM2.5 with
full term low birth weight at different spatial scales. Environmental Research, 134, 427–434.
https://doi.org/10.1016/j.envres.2014.05.034
Hawsawi, A. M., Bryant, L. O., & Goodfellow, L. T. (2015). Association Between Exposure to
Secondhand Smoke During Pregnancy and Low Birthweight: A Narrative Review. Respiratory
Care, 60(1), 135–140. https://doi.org/10.4187/respcare.02798
Jaakkola, M. S., & Jaakkola, J. J. (1997). Assessment of exposure to environmental tobacco smoke.
European Respiratory Journal, 10(10), 2384–2397.
Jornayvaz, F. R., Vollenweider, P., Bochud, M., Mooser, V., Waeber, G., & Marques-Vidal, P. (2016).
Low birth weight leads to obesity, diabetes and increased leptin levels in adults: The CoLaus
study. Cardiovascular Diabetology, 15(1), 73. https://doi.org/10.1186/s12933-016-0389-2
Kioumourtzoglou, M.-A., Spiegelman, D., Szpiro, A. A., Sheppard, L., Kaufman, J. D., Yanosky, J. D.,
Williams, R., Laden, F., Hong, B., & Suh, H. (2014). Exposure measurement error in PM2.5
health effects studies: A pooled analysis of eight personal exposure validation studies.
Environmental Health, 13, 2. https://doi.org/10.1186/1476-069X-13-2
Laurent, O., Hu, J., Li, L., Kleeman, M. J., Bartell, S. M., Cockburn, M., Escobedo, L., & Wu, J. (2016).
Low birth weight and air pollution in California: Which sources and components drive the risk?
Environment International, 92–93, 471–477. https://doi.org/10.1016/j.envint.2016.04.034
Li, X., Huang, S., Jiao, A., Yang, X., Yun, J., Wang, Y., Xue, X., Chu, Y., Liu, F., Liu, Y., Ren, M.,
Chen, X., Li, N., Lu, Y., Mao, Z., Tian, L., & Xiang, H. (2017). Association between ambient
fine particulate matter and preterm birth or term low birth weight: An updated systematic review
and meta-analysis. Environmental Pollution, 227, 596–605.
https://doi.org/10.1016/j.envpol.2017.03.055
Li, Z., Tang, Y., Song, X., Lazar, L., Li, Z., & Zhao, J. (2019). Impact of ambient PM2.5 on adverse birth
outcome and potential molecular mechanism. Ecotoxicology and Environmental Safety, 169,
248–254. https://doi.org/10.1016/j.ecoenv.2018.10.109
Madsen, C., Gehring, U., Erik Walker, S., Brunekreef, B., Stigum, H., Næss, Ø., & Nafstad, P. (2010).
Ambient air pollution exposure, residential mobility and term birth weight in Oslo, Norway.
Environmental Research, 110(4), 363–371. https://doi.org/10.1016/j.envres.2010.02.005
Moore, S. E. (2017). Early-Life Nutritional Programming of Health and Disease in The Gambia. Annals
of Nutrition and Metabolism, 70(3), 179–183. https://doi.org/10.1159/000456555
O’Connor, T. Z., Holford, T. R., Leaderer, B. P., Hammond, S. K., & Bracken, M. B. (1995).
Measurement of Exposure to Environmental Tobacco Smoke in Pregnant Women. American
Journal of Epidemiology, 142(12), 1315–1321.
https://doi.org/10.1093/oxfordjournals.aje.a117599
Pérez-Ríos, M., Schiaffino, A., López, M. J., Nebot, M., Galán, I., Fu, M., Martínez-Sánchez, J. M.,
Moncada, A., Montes, A., Ariza, C., & Fernández, E. (2013). Questionnaire-based second-hand
smoke assessment in adults. European Journal of Public Health, 23(5), 763–767.
https://doi.org/10.1093/eurpub/cks069
Pogodina, C., Brunner Huber, L. R., Racine, E. F., & Platonova, E. (2009). Smoke-Free Homes for
Smoke-Free Babies: The Role of Residential Environmental Tobacco Smoke on Low Birth
Weight. Journal of Community Health, 34(5), 376–382. https://doi.org/10.1007/s10900-009-
9169-1

9
Prüss-Üstün, A., Wolf, J., Corvalán, C. F., Bos, R., & Neira, M. P. (2016). Preventing disease through
healthy environments: A global assessment of the burden of disease from environmental risks.
World Health Organization. https://apps.who.int/iris/handle/10665/204585
Rich, D. Q., Liu, K., Zhang, J., Thurston, S. W., Stevens, T. P., Pan, Y., Kane, C., Weinberger, B.,
Ohman-Strickland, P., Woodruff, T. J., Duan, X., Assibey-Mensah, V., & Zhang, J. (2015).
Differences in Birth Weight Associated with the 2008 Beijing Olympics Air Pollution Reduction:
Results from a Natural Experiment. Environmental Health Perspectives, 123(9), 880–887.
https://doi.org/10.1289/ehp.1408795
Savitz, D. A., Bobb, J. F., Carr, J. L., Clougherty, J. E., Dominici, F., Elston, B., Ito, K., Ross, Z., Yee,
M., & Matte, T. D. (2014). Ambient Fine Particulate Matter, Nitrogen Dioxide, and Term Birth
Weight in New York, New York. American Journal of Epidemiology, 179(4), 457–466.
https://doi.org/10.1093/aje/kwt268
Schellong, K., Schulz, S., Harder, T., & Plagemann, A. (2012). Birth weight and long-term overweight
risk: Systematic review and a meta-analysis including 643,902 persons from 66 studies and 26
countries globally. PloS One, 7(10), e47776. https://doi.org/10.1371/journal.pone.0047776
Schembari, A., de Hoogh, K., Pedersen, M., Dadvand, P., Martinez, D., Hoek, G., Petherick, E. S.,
Wright, J., & Nieuwenhuijsen, M. J. (2015). Ambient Air Pollution and Newborn Size and
Adiposity at Birth: Differences by Maternal Ethnicity (the Born in Bradford Study Cohort).
Environmental Health Perspectives, 123(11), 1208–1215. https://doi.org/10.1289/ehp.1408675
Sun, X., Luo, X., Zhao, C., Zhang, B., Tao, J., Yang, Z., Ma, W., & Liu, T. (2016). The associations
between birth weight and exposure to fine particulate matter (PM2.5) and its chemical
constituents during pregnancy: A meta-analysis. Environmental Pollution, 211, 38–47.
https://doi.org/10.1016/j.envpol.2015.12.022
Tsai, J. (2018). Exposure to Secondhand Smoke Among Nonsmokers—United States, 1988–2014.
MMWR. Morbidity and Mortality Weekly Report, 67.
https://doi.org/10.15585/mmwr.mm6748a3
Upadhyay, R. P., Naik, G., Choudhary, T. S., Chowdhury, R., Taneja, S., Bhandari, N., Martines, J. C.,
Bahl, R., & Bhan, M. K. (2019). Cognitive and motor outcomes in children born low birth
weight: A systematic review and meta-analysis of studies from South Asia. BMC Pediatrics,
19(1), 35. https://doi.org/10.1186/s12887-019-1408-8
Vilanova, C. S., Hirakata, V. N., de Souza Buriol, V. C., Nunes, M., Goldani, M. Z., & da Silva, C. H.
(2019). The relationship between the different low birth weight strata of newborns with infant
mortality and the influence of the main health determinants in the extreme south of Brazil.
Population Health Metrics, 17(1), 15. https://doi.org/10.1186/s12963-019-0195-7
Watkins, W. J., Kotecha, S. J., & Kotecha, S. (2016). All-Cause Mortality of Low Birthweight Infants in
Infancy, Childhood, and Adolescence: Population Study of England and Wales. PLoS Medicine,
13(5). https://doi.org/10.1371/journal.pmed.1002018
Whitaker, A. H., Feldman, J. F., Lorenz, J. M., Shen, S., McNicholas, F., Nieto, M., McCulloch, D.,
Pinto-Martin, J. A., & Paneth, N. (2006). Motor and Cognitive Outcomes in Nondisabled Low-
Birth-Weight Adolescents: Early Determinants. Archives of Pediatrics & Adolescent Medicine,
160(10), 1040. https://doi.org/10.1001/archpedi.160.10.1040
Xiao, X., Li, Y., Song, X., Xu, Q., Yang, S., Wu, J., & Seto, E. (2018). Discrepancy between Self-
Reported and Urine Cotinine-Verified Environmental Tobacco Smoke Exposure among Rural
Pregnant Women in China. International Journal of Environmental Research and Public Health,
15(7), 1499. https://doi.org/10.3390/ijerph15071499
Yang, Y., Pun, V. C., Sun, S., Lin, H., Mason, T. G., & Qiu, H. (2018). Particulate matter components
and health: A literature review on exposure assessment. Journal of Public Health and Emergency,
2(3), Article 3. https://jphe.amegroups.com/article/view/4438

10
CHAPTER 1
BACKGROUND AND REVIEW
MADRES STUDY
Data for the dissertation comes from a personal particulate matter with aerodynamic diameter less
than 2.5µm (PM 2.5) exposure monitoring study nested with the Maternal and Developmental Risks from
Environmental and Social Stressors Center for Environmental Health Disparities (MADRES) at the
University of Southern of California (USC). A full explanation of MADRES pregnancy cohort is found
elsewhere (Bastain et al., 2019). Briefly, MADRES is a prospective cohort study created to investigate the
cumulative impact of environmental pollutants and psychosocial, behavioral, and built environmental risk
factors on maternal and infant health outcomes, in a health and environmental pollution disparities
population in Los Angeles, CA. This cohort consists of ~900 predominantly Hispanic, lower-income
pregnant women, who were enrolled through community partnerships with four prenatal care providers
beginning November 2015. Participation eligibility included the following criteria: 1) less than 30 weeks
gestation at recruitment, 2) over 18 years of age, 3) speak English or Spanish, 4) HIV negative, 5) single
gestation, 6) no physical, mental, or cognitive disability that hinder study participant and proving informed
consent.
The personal PM 2.5 exposure monitoring study took place in the 3
rd
trimester of pregnancy with 214
participants and used a 48-hour integrated custom sampling approach between October 2016 and February
2020. Study participants wore a personal sampling apparatus consisting of a crossbody purse containing a
Gilian Plus Datalogging Pump (Sensidyne Inc., Clearwater, FL), which was programmed to start at
midnight (the following day) and actively sample at a 50% cycle and flow rate of 1.8 liters per minute
(LPM). A Harvard PM 2.5 personal exposure monitor (PEM) with a 37mm Pall Teflo filter was connected to
the pump, with a sampling inlet located near the shoulder around the breathing zone. Participants were
provided with instructions and a demonstration about correct usage and were requested to wear the sampling
device as much as possible while going through normal daily activities to obtain the most realistic
assessment as possible. A limited number of activities were considered exceptions including, driving,

11
showering, and sleeping, with the added request that participants protect the device from water, heat,
humidity, pets, and children. When not wearing the device, participants were asked to keep the device close
by, such as on a passenger seat or bedside table. At the end of the 48-hour study period, the sampling pump
was programmed to shut off, with a staff member collecting the device the following day and conducting a
brief exit survey. When the sampling devices arrived at the USC Exposure Analytics lab, they were handled
by trained staff. Pump data were downloaded, checked for errors, and securely stored. Filters were removed
from the PEMs, allowed to equilibrate within a dedicated chamber, and gravimetrically weighed in
temperature and relative humidity-controlled glove box using an MT-5 calibrated microbalance to obtain
PM 2.5 mass concentration.
MADRES collects data at several time points across pregnancy, ranging from recruitment, each
trimester, birth, and postnatal. Data collected included maternal demographics, pregnancy and birth
outcomes, medical histories, lifestyle factors, and meteorology. Due to the personal exposure monitoring
study being conducted in the 3
rd
trimester, the exit survey and 3
rd
trimester questionnaires administered by
trained staff members were used, when possible, to have the most time-sensitive data to match the exposure
monitoring data used.
HEALTH DISPARITIES
Health disparities are preventable differences in the incidence, prevalence, and burden of disease
and mortality by socially disadvantaged populations (National Academies of Sciences, 2017), with
disparities for several outcomes seen in the earliest years of life. There are racial/ethnicity differences in
obesity (Karnes et al., 2021), childhood asthma (Banta et al., 2021), and several cancers, including but not
limited to, testicular cancer (Y. Li et al., 2020) and lung cancer in adults (Spracklen et al., 2014), and
leukemia in children (Oksuzyan et al., 2015). Additionally, there are racial/ethnicity differences observed
in the prevalence of type-2 diabetes, with non-Hispanic Blacks and Hispanics experiencing the greatest
burden of disease of 12.7% and 12.1%, respectively, compared to just 7.4% of non-Hispanic Whites (CDC
Newsroom, 2016). Furthermore, prediabetes, a condition that if untreated will likely lead to type-2 diabetes
within 5 years, was also higher for these two groups compared to non-Hispanic Whites, with an age-

12
standardized prevalence of 35.3%, 32.0%, and 31.0%, for Hispanics, non-Hispanic Blacks, and non-
Hispanic Whites, respectively (Y. Zhu et al., 2019). Such disparities may also exist in terms of
complications from diabetes (Spanakis & Golden, 2013), potentially increasing the burden on health
disparities populations.
Health disparities are prominent and persistent and can be seen throughout the whole life of
individuals. For example, Gee et al., (2013) reported that in Northern California, race/ethnicity differences
in obesity were present as early as 2-5 years, with the proportions of individuals that were obese highest in
Hispanics at 16.0% (95% CI: 15.4, 16.6), followed by Blacks 12.2% (11.6, 12.9), Asians 8.1% (7.8, 8.4),
and Whites 7.7% (7.4, 7.9), respectively. Additionally, these disparities were seen even after adjustment
for age and sex, with race/ethnicity differences in body mass index (BMI) Z-scores observed at 2 years with
Hispanics having a BMI Z-score of 0.66 (SE: 0.11), Blacks 0.58 (0.12), Whites 0.44 (0.07), and Asians
0.28 (0.12), respectively (Isong et al., 2018). The racial differences persist through preschool and
kindergarten (Gee et al., 2013; C. M. Hales, 2020; Isong et al., 2018), and are not completely explained by
socioeconomic status (Rossen, 2014).
THE IMPACT OF BIRTHWEIGHT ON HEALTH
Birthweight is a birth outcome that has garnered a great deal of academic interest due to evidence
that birthweight impacts both short- and long-term health (Belbasis et al., 2016). Most research with
birthweight has concentrated on low birthweight (LBW), an outcome defined as a birth below 2,500 grams
(g) in weight, which is observed in around 8.3% of newborns (Martin et al., 2018), and also varies by
race/ethnicity in this United States (U.S.), with non-Hispanic Blacks, Hispanics, and non-Hispanic Whites
having 11.4%, 6.1%, and 5.2% of LBW births, respectively (Womack et al., 2018). However, the
prevalence of high birthweight (HBW; macrosomia), a birth above 4,000g, is around 7.8% nationally, and
may also present an increased risk of negative health outcomes. Additionally, neighborhood deprivation
was positively associated with small- and large-for gestational age (Wentz et al., 2014), highlighting how
socioeconomic status may also impact birth outcomes related to fetal growth.

13
Interestingly, LBW is associated with several adverse outcomes throughout life. For example, LBW
has been linked to obesity in childhood and adult obesity (D. J. P. Barker, 2004; C. Chen et al., 2019)
through intrauterine growth restriction (IUGR) that causes fetal adaptions, which may adversely impact
lifetime health (discussed in more detail below). However, a U-shaped relationship may be present for some
metabolic conditions, with studies showing that both very high and low birth weight are associated with
higher BMI and adiposity in later life (Qiao et al., 2015; Stansfield et al., 2016), with both thinness and
severe obesity also being observed in children between 3-12 years (C. Chen et al., 2019). Additionally,
there was also a U-shaped association with type-2 diabetes, with a 47% (95% CI: 1.26, 1.72) and 36% (1.07,
1.73) increased risk (vs. “normal” (≥ 2,500g)) for both those born <2,500g and >4,000g, respectively
(Harder et al., 2007).
The impact of LBW and HBW extend beyond metabolic health conditions. Research found that
infant mortality was associated with being LBW (Watkins et al., 2016), which seems especially true for
very LBW (<1,500g) (Vilanova et al., 2019). Others found LBW was linked to cognitive development
(Upadhyay et al., 2019; Whitaker et al., 2006), and cardiovascular disease (Huxley et al., 2007; Risnes et
al., 2011; Smith et al., 2016). Numerous cancers are associated with both LBW but also HBW (Belbasis et
al., 2016). For example, a meta-analysis found a 35% (95% CI: 1.25, 1.48) increased risk of overall
childhood leukemia for HBW and a 50% (95% CI: 1.05, 2.13) increased risk of acute myeloid leukemia for
LBW (Caughey & Michels, 2009).
The impact of birthweight on overall health, both short- and long-term, has drawn a lot of interest
in factors that contribute to changes in birthweight. Environmental factors are a potential causal exposure,
especially due to the racial differences explored above in health outcomes and the fact that Hispanic and
Black populations continue to face a greater burden of several environmental exposures, including air
pollution, such as particulate matter and its components (Bell & Ebisu, 2012; Tessum et al., 2021). Overall,
these findings highlight that racial disparities exist in both health risks and environmental exposures and
may in fact be contributing to a disproportionate burden of negative health outcomes in certain racial

14
groups. Therefore, this dissertation evaluated whether PM 2.5 (and its major sources and components),
measured with less biased exposure assessment tools, was associated with birthweight.
BIOMECHANISM OF LOW BIRTHWEIGHT IMPACTS ON HEALTH
How birthweight affects short- and long-term health is still a debated issue but physical changes or
adaptions during fetal development seem to be important factors that has garnered a great deal of research.
The “developmental origin of chronic adult disease” is a well-established hypothesis for how birthweight,
specifically LBW, impacts health in later life (D. J. P. Barker, 2004). The concept is predicated on
assumptions that environmental insults acting in the earliest stages of life (often in the prenatal period),
have profound effects on health through permanent changes in physiology and metabolism (Boo & Harding,
2006). Barker and colleagues developed this hypothesis on observations that regions of the United Kingdom
that had the highest levels of infant mortality, also had the highest rates of coronary heart disease, suggesting
common causal mechanisms (D. J. Barker & Osmond, 1986). They then expanded this concept to metabolic
and associated disorders including diabetes, hypertension, and stroke, through adverse effects on glucose-
insulin metabolism from poor fetal growth (D. J. P. Barker, 2004; C. N. Hales & Barker, 2001). It is now
widely accepted that a “hostile intrauterine environment,” which is effectively a uterus with less than
optimum conditions for fetus growth and development, from environmental insults cause fetal adaptions in
the form of hemodynamic, metabolic, epigenetic alterations, and hormonal adjustments, in an attempt to
manage the ill-effects of the adverse intrauterine environment but may increase the risk of a host of ailments
(Hoffman et al., 2017; Priante et al., 2019).
Growth restriction is thought to have several causes, including but not limited to, fetal factors,
including chromosomal abnormalities, genetic syndromes, and intrauterine infections, and maternal factors,
including, genetics, nutrition, preeclampsia, and metabolic disorders, as well as, placental factors, such as
inadequate blood flow and decreased nutrition delivery (Lin & Santolaya-Forgas, 1998; Nardozza et al.,
2017). Figure 1.1 lists some maternal, placental, and fetal factors from the literature that are likely
important factors for how environmental exposures contribute to alterations in fetal growth (Kamai et al.,
2019; Stapleton, 2016).

15
Figure 1.1. Maternal, placental, and fetal factors in fetal growth impacted by environmental exposures.

Notes: Adapted from Kamai et al,. 2010.
Such factors in turn lead to the development of insulin resistance, visceral obesity, glucose
intolerance, increased leptin levels, and higher leptin to fat mass ratio, which subsequently leads to a greater
risk of obesity, type-2 diabetes, and metabolic syndrome in adulthood (Barnes & Ozanne, 2011; C. N.
Hales, 1997; Jornayvaz et al., 2016; Morrison et al., 2010; Visentin et al., 2014). In the case of type-2
diabetes, it has been suggested that because only two of the 45 known type-2 diabetes susceptibility genes
are associated with LBW that any association between LBW and type-2 diabetes is largely from non-genetic
sources (Vaag et al., 2012). In the case of LBW and its possible link to obesity, there is an increase in the
abundance of insulin receptors and the insulin signaling pathway of IUGR infants, possibly due to an
adaption to poor nutrient supply (Morrison et al., 2010). This in turn creates conditions for excessive weight
gain in early post-natal life.
However, this may differ by sex, as research suggests that a U-shaped association between
birthweight and odds of childhood obesity (9-11 years of age) existed for boys but was reasonably linear
for girls. For example, in boys, and using 2,500-2,999g as the reference group, those born < 2,500g (LBW)
had an odds ratio OR (95% CI) of 1.37 (0.75-2.50) of developing obesity in childhood, compared to an OR

16
of 1.28 (0.88-1.88) for those born 3,000-3,499g, and finally, for those ≥ 4000g, an OR of 1.77 (1.12-2.82)
(Qiao et al., 2015). Furthermore, birthweight is inversely associated with blood pressure, obesity, and
abdominal obesity among 10-13-year-old school children (M et al., 2013), while others found a U-shaped
relationship with systolic blood pressure (C et al., 2019). Additionally, birthweight has also been shown to
impact fat intake in school-aged boys, with fetal programming of “homeostatic and/or hedonic pathways”
influencing food preferences (Ar et al., 2018). This highlights that it may not necessarily be just the adverse
effects of LBW that prime individuals for future health issues, but that either shared or separate pathways
for HBW, may also place individuals at excess risk.
In-utero PM 2.5 exposure is hypothesized to contribute to a hostile intrauterine environment by
causing many of the biomechanistic pathways discussed above, such as oxidative stress, DNA methylation
changes, mitochondrial DNA content alteration, and endocrine disruption (Clemente et al., 2016; Z. Li et
al., 2019). In-utero PM 2.5 exposure may also affect gene expression and the interplay of epigenetic and
environmental factors, placing individuals at greater risk of short- and long-term health conditions,
including asthma (Korton et al., 2017). Black carbon (BC), a component of PM 2.5 that is derived from
incomplete combustion of fuels and diesel-burning soot emissions, was found accumulated on the fetal side
of the human placenta, representing a potential mechanism for adverse health effects (Bové et al., 2019),
possibly through intrauterine inflammation, which was positively associated with in utero PM 2.5 exposure
(Nachman et al., 2016). In general, PM 2.5 exposure also decreases red blood count, monocyte count, and
increases lymphocyte count, providing further possible routes for adverse health effects (Z. Li et al., 2021).
Additionally, sources of PM 2.5 such as prenatal smoking may impact telomere length (Wei et al., 2020),
which may lead to gastric cancers, diabetes, and Alzheimer’s diseases (L. Smith et al., 2019).
Taken together, it appears that the effect of PM 2.5 on birthweight (predominately LBW but
potentially also HBW), possibly through fetal reprogramming and hostile intrauterine environments, makes
individuals more susceptible to a host of adverse health outcomes, including obesity and metabolic-related
conditions both in early and later life. Understanding the relationship of personal PM 2.5 exposure during

17
pregnancy with birthweight is an important health question due to the persistency of this environmental
exposure, and the potential impact it may be having on both early and later life disease.
OUTDOOR PARTICULATE MATTER AIR POLLUTION
Within the United States (US), the Environmental Protection Agency (EPA) sets federal regulations
called the National Ambient Air Quality Standards (NAAQS) to control levels of ambient air pollutants
(outdoor or background concentrations), including PM 2.5. The current U.S. standard for annual average
PM 2.5 ambient concentrations is 12 μg/m
3
(Giannadaki et al., 2016). However, even with regulatory
mechanisms in place and below current standards, air pollution continues to have a negative impact on
health (Bell & Ebisu, 2012).
PM 2.5 is an air pollutant that is a mixture of solid particles and components, including carbons,
metals, and crustal elements, and can originate from several sources including fuel oils, secondhand
smoking, industrial processes, wildfires, and traffic (Adams et al., 2015). PM 2.5 can be directly emitted in
the form of primary particles from a given source or formed as secondary particles through chemical
reactions in the atmosphere (Hodan & Barnard, 2004). PM 2.5 is roughly 30 times smaller than a human hair,
often making individuals unaware of exposure, but also allowing for particles to be inhaled into the lungs
and reach the lower respiratory tract and alveolar region (US EPA, 2016).
Sources of PM 2.5 can vary in size distribution and chemical composition. Each source may have
unique signatures or high-loading component(s) that may act individually or synergistically. For example,
natural marine aerosol sources, such as aged and/or fresh sea salt, have high sodium, chlorine, and
magnesium, while vehicular emissions are often high in black carbon (BC), zinc, barium, and iron (Banerjee
et al., 2015). Notably, the components are not equally distributed by region and by socio-demographics,
including race/ethnicity and education. For example, Basu et al., (2014) highlighted variation in
concentration of PM 2.5 and its components across California, highlighting potential heterogeneity in PM 2.5
sources with different risk levels. For example, between 2000 and 2006, mean (SD) PM 2.5 concentration
during the full gestational period was 25.3 (2.5) μg/m
3
in Riverside and 13.1 (1.6) μg/m
3
in Simi Valley,
nitrate concentration was 9.5 (1.8) μg/m
3
in Riverside and just 2.5 (0.8) μg/m
3
in Sacramento, and

18
ammonium was 3.7 (0.7) μg/m
3
in Riverside and 0.9 (0.2) μg/m
3
in San Jose. Additionally, Bell and Ebisu
(2012), found that in California, Hispanics and non-Hispanic Blacks had elevated exposures of 13 out of
the 14 elements they measured compared to non-Hispanic Whites, including aluminum, nickel, calcium,
chlorine, vanadium, titanium, nitrate, and zinc. They also found that individuals who did not graduate high
school had higher residential concentrations of all 14 elements compared to college graduates. Combined
this suggests the groups that face a higher burden of health disparities are also exposed to a greater burden
of environmental exposures.
Exposure to outdoor PM 2.5 can be assessed at different spatial scales, which impacts the levels of
measurement error in the exposure estimate (Carroll, 2005; Zeger et al., 2000). Measurements from
regulatory monitoring sites are often used to predict residential exposure to ambient (general background)
or outdoor (at the residence) PM 2.5. More recent modeling methods including spatiotemporal modeling,
typically incorporating remote sensing, and/or geospatial data (Dadvand et al., 2013; Ebisu et al., 2014;
Gray et al., 2010; Harris et al., 2014). These methods provide more accurate estimates of outdoor residential
PM 2.5 and offer several advantages over ambient approaches, including capturing spatial variability and
temporal changes in PM 2.5 concentrations, which combined may reduce error (Di et al., 2019). While these
approaches have led to great advances towards reducing exposure measurement error in PM 2.5 assessment,
error remains in terms of estimating personal exposure to PM 2.5 of outdoor origin, which likely results in
biased effect estimates and attenuated power to detect health effects depending on the type or direction of
the error (Carroll, 2005; Kioumourtzoglou et al., 2014; Zeger et al., 2000). Additionally, researchers have
raised concerns about the spatial variability of PM 2.5 sources and components, which likely leads to
exposure misclassification from the use of ambient monitors to estimate individual exposure (Bell et al.,
2011). In general, these methods introduce exposure error because they ignore the fact that individuals
spend the majority of their time indoors, and their “true” personal exposure to PM 2.5 of outdoor origin is a
result of the infiltration efficiency of PM 2.5 indoors and time-activity patterns or where they spend their time
and what they do, which are collectively most accurately captured by personal monitoring (Gray et al.,
2011).

19
PERSONAL PM2.5 EXPOSURE
Total personal PM 2.5 exposure encompasses contributions from indoor, outdoor, and personal
activity-related sources. It is a measure that integrates or accounts for the infiltration of outdoor air pollution
indoors (or into the personal cloud), as well as indoor sources and time-activity patterns and behaviors of
individuals. As discussed above, traditional approaches to PM 2.5 exposure assessment incorporate ambient
monitoring, remote sensing, and other geospatial data to predict outdoor PM 2.5 concentrations at the
residential location with the assumption that this outdoor estimate represents or captures personal exposure
to outdoor-origin air pollution. Therefore, they do not estimate the total personal PM 2.5 an individual is
exposed to, which itself is a separate but equally important exposure in need of research on health impacts.
PM 2.5 of outdoor origin, or the combination of various outdoor sources of PM 2.5, may make up a small
fraction of an individual’s total exposure to PM 2.5 and this might also vary across contexts and geography
(Wallace et al., 2003). However, PM 2.5 of outdoor origin is the fraction of PM 2.5 exposure that is regulated,
so it is an important exposure to evaluate to support science risk assessments for setting air quality
standards, albeit a different question to the effects of total personal PM 2.5 exposure on health. Also, Avery
et al., (2010) found that the adequacy of ambient or outdoor PM 2.5 as a proxy or surrogate for total personal
exposure varied significantly with correlation coefficients ranging from 0.09-0.83, possibly due to the
characteristics of studies, environments, and among participants. Interestingly, researchers in Denmark
found that total personal PM 2.5 exposure appears more closely related to oxidative stress biomarkers in the
blood than ambient PM 2.5 concentrations, with personal carbon black producing the greatest increase in
plasma proteins (Sørensen et al., 2003). This is a potential mechanism for how PM 2.5 impacts birthweight
(discussed above), highlighting that both measurement error and exposure type considerations are important
in understanding the health effects of PM 2.5 exposure.
The “gold standard” of individual-level external PM 2.5 exposure assessment is personal monitoring
that directly (or as close as possible) measures concentrations of air pollutants in the breathing zone of
individual participants as they go about their everyday lives. This approach alleviates much of the
measurement error and bias introduced using outdoor air pollution assessment methods as it captures

20
outdoor, indoor, and personal influences on PM 2.5. Briefly, it accounts for the infiltration of outdoor sources
into the home or other indoor spaces where participants might spend most of their time, impacts of indoor
sources and behaviors that emit or affect PM 2.5, and also the time-activity patterns of individuals (or the
time they spend in different microenvironments such as in-transit, indoor, or outdoor).
However, this approach is not without its limitations. It is often very logistically taxing and
expensive to conduct such an exposure measurement study at scale, especially in pregnancy studies where
the burden of wearing the device may be greater, and as such, only a limited number of studies have
attempted to assess personal exposure in pregnancy studies this way. The majority of published literature
has done so to investigate the health effects of exposure to toxic polyaromatic hydrocarbons which are a
very specific class of chemicals emitted from incomplete fuel combustion and especially diesel (Choi et al.,
2012, 2008; Rundle et al., 2012).
ASSOCIATION BETWEEN PRENATAL EXPOSURE TO PM2.5 AND BIRTHWEIGHT
Several studies have investigated the effect of PM 2.5 exposure during pregnancy on birthweight and
risk of LBW with a modest but significant finding that generally supports the notion that ambient or outdoor
PM 2.5 leads to lower birthweight and an increased risk of being a LBW birth (Pedersen et al., 2013; Savitz
et al., 2014; Schembari et al., 2015). For example, a meta-analysis found that there was a decrease of 22.2g
(95% CI: -37.9, -6.4) per 10 μg/m
3
of PM 2.5 exposure during the entire pregnancy (Lamichhane et al., 2015).
Another found a statistically significant 15.9g reduction in birthweight (95% CI: -26.8, -5.0) and a 9% (1.0,
1.2) increased risk of LBW per 10 μg/m
3
increase in PM 2.5 also through the entire pregnancy (Sun et al.,
2016). Additionally, the effect of PM 2.5 on birthweight seems to vary by time period with the 3
rd
trimester
of pregnancy appearing to have the greatest impact (Dadvand et al., 2014; Savitz et al., 2014; Schembari et
al., 2015), which is also the window of time where most fetal weight gain occurs (Kiserud et al., 2018). A
review found a decrease of 6.6g (95% CI: -13.7, 0.4) in the 1
st
trimester, 8.0g (-14.5, -1.5) decrease in the
2
nd
trimester, and 14.9g (-21.7, -8.1) decrease in the 3
rd
trimester per 10 μg/m
3
of PM 2.5, respectively (X.
Zhu et al., 2015). Yuan et al (2020) also found that the critical window of in utero exposure to PM 2.5 in the

21
Shanghai birth cohort study was 31-34 gestational weeks for a reduction in birthweight, and 38 – 42 weeks
for an increased risk of LBW (Yuan et al., 2020).
However, heterogeneous results are present in the literature (Stieb et al., 2012; Sun et al., 2016; X.
Zhu et al., 2015), likely due to measurement error introduced by estimating personal exposure of outdoor
origin from ambient data (Kioumourtzoglou et al., 2014). Few studies have used personal exposure
assessment to evaluate the effect of PM 2.5 on birthweight. A study based in Poland analyzed total personal
PM 2.5 using personal monitoring over a 48-hour period in a cohort of female non-smokers and found that
there was a 97.2g (95% CI -201.0, 6.6) decrease in birthweight per increase in 30 μg/m³ in PM 2.5
(Jedrychowski et al., 2004, 2009). Additionally, this decrease in birthweight as a result of increasing
personal PM 2.5 differed by sex, with an average decrease of 189g for males compared to just 17g for females
(Jedrychowski et al., 2009). However, this study was conducted in a racially homogenous population, and
in the 2
nd
trimester, rather than in the 3
rd
trimester of pregnancy where most of the fetal growth takes place.
Additionally, there are regional differences in the relationship between ambient or outdoor PM 2.5
and risk of lower birthweight. For example, an association between PM 2.5 and decreased birthweight was
found in Poland (Wojtyla et al., 2020), but not in Norway (Madsen et al., 2010). Additionally, an increased
risk of LBW from PM 2.5 exposure was found in California (Basu et al., 2014), but not in North Carolina
(Gray et al., 2014). This may be due to differences in measurement assessment methodology (Richmond-
Bryant & Long, 2020), but it is also likely that regional differences in PM 2.5 components with differing
toxicities may be driving these heterogeneous findings. Between 2000 and 2006, a pregnancy study found
that nitrate was a dominant PM 2.5 component in Los Angeles, while the concentration for ambient nitrate
in other places in California, such as San Jose and Sacramento, was noticeably lower (Basu et al., 2014).
Similarly, the same researchers found vanadium and nickel concentrations (components linked to a
reduction in birthweight) in Los Angeles were over double that of Sacramento and San Jose, respectively
(Basu et al., 2014).
Furthermore, a multi-country meta-analysis revealed that regions with high PM 2.5/PM 10 ratios,
indicative of regions heavily impacted by sources of primary combustion (for example, Los Angeles, CA),

22
had greater associations with LBW (Dadvand et al., 2013). Additionally, within Los Angeles County,
significant spatial heterogeneity in PM 2.5 and LBW was observed, with the highest and most confident
effects seen in the Central and Southern LA County tracts which constitute areas close to the MADRES
catchment area (Coker et al., 2015). This heterogeneity or discrepancy in health findings is likely due to the
variable chemical composition or makeup of PM 2.5 in these different regions. Given that PM 2.5 is a mixture
of organic and inorganic components (discussed in more detail below), including carbons, metals, and
crustal elements, its toxicity has been shown to vary based on the chemical composition (Laurent et al.,
2014, 2016; Ng et al., 2017).
ASSOCIATION BETWEEN OUTDOOR PM2.5 SOURCES AND COMPONENTS AND
BIRTHWEIGHT
While many studies have looked at the association between outdoor PM 2.5 on birthweight, the effect
of specific PM 2.5 components have received much less attention. This may be due to the more limited
measurement of these components at ambient air pollution monitors in some regions (for example, usually
measured every third or sixth day at one station in a large area), but examining these specific components
rather than ambient PM 2.5 mass may more accurately capture the source-specific risks that individuals face
regarding reductions in birthweight. It may also provide insight into the biological mechanisms unpinning
this association and explain some of the racial and socioeconomic status disparities seen, especially
considering Hispanics and Blacks face a greater burden of exposure to some of these components compared
to non-Hispanic Whites (Bell & Ebisu, 2012). Additionally, because PM 2.5 levels, sources, and chemical
components differ between the East and West Coasts and even within urban areas at neighborhood scales,
studying different sources of PM 2.5 and its respective components may explain some of the heterogeneity
seen in PM 2.5 health analyses.
For example, Basu et al., (2014) observed a reduction of 6g of birthweight (95% CI: -9,-4) per 1
IQR (7.56 μg/m
3
) increase of PM 2.5 mass in California. However, interestingly, they also found that PM 2.5
components associated with traffic, sulfur, and metals sources (sulfur, vanadium, sulfate, iron, and

23
elemental carbon), had a more toxic effect on birthweight. For example, while potassium and calcium had
a minimal effect on birthweight of -2g (95% CI: -3, -1) and 1g (0, 3), respectively, vanadium, sulfur, iron,
and elemental carbon were associated with much greater reductions in birthweight of -32g (-38, -27), -29
(-33, -25), -21 (-24, -18), and -16 (-19,-14), respectively (Basu et al., 2014).
While having a limited number of studies to work with for specific PM 2.5 components, a meta-
analysis of PM 2.5 constituents was attempted by Sun et al., (2016) and they similarly found that several
components saw a negative change in birthweight of -38.3g (95% CI: -45.3, -31.2) for vanadium, titanium
-21.5g (-24.3, -18.7), nickel -16.2g (-21.0, -11.3), elemental carbon -16.1g (-30.0, -2.2), silicon -25.2g (-
54.4, 3.9), and sulfur -32.2g (-72.5, 8.1). Additionally, they found that the risk of LBW (OR (95% CI)
increased for an increase in several components, including titanium 1.2 (1.0, 1.3), vanadium 1.1 (1.0, 1.3),
elemental carbon 1.1 (1.0, 1.2), silicon 1.4 (1.0, 1.9), and sulfur 1.2 (1.1, 1.4), highlighting that specific
components of PM 2.5 may be explaining or driving some of the risks associated with PM 2.5 mass.
However, while some studies highlight particular components as potentially playing a larger role
on the toxic effect of PM 2.5 mass on birthweight, others show similar heterogeneity among their results as
seen with PM 2.5 mass more broadly. For example, Bell et al., (2010) found a statistically significant
increased risk of LBW for aluminum, while Basu et al., (2014) found no association. Laurent et al., (2016)
found that elemental carbon was not associated with LBW, while Sun et al., (2016) did.
Possible reasons for the heterogeneity may reflect the measurement error witnessed in air pollution
assessment discussed above, where inaccurate ambient monitor data and/or assessment methodology fails
to capture an individual’s “true” personal exposure to these components, resulting in attenuated effects.
This may be increased when studying individual components of PM 2.5, which may be more spatially
variable than PM 2.5 mass (Bell et al., 2011). Another possible reason for the differing results is that specific
sources of PM 2.5 when present may be driving the greater risk of LBW including, wildfire smoke, oil
combustion, second-hand smoke, and traffic. In the case of traffic sources of PM 2.5, both exposure to on-
road gasoline and diesel emissions have been found to be associated with birthweight (Bell et al., 2010;
Laurent et al., 2016; Wilhelm et al., 2012). Additionally, a recent meta-analysis found a significant positive

24
association between traffic density and LBW (Wang et al., 2020), highlighting that there may be something
about traffic sources or the traffic-related mixture that has a greater toxic effect on birthweight.
Finally, within the 3
rd
trimester of study, which is the trimester used in this analysis, the effect of
PM 2.5 components and/or sources on birthweight were shown to differ by race, with the negative effect of
sulfate on birthweight greatest in Blacks, with a 11.7g decrease in birthweight (95% CI: -22.9, -0.6), a
decrease of 9.4g (-22.0, 3.2) in Hispanics, and lowest in Whites 5.1 (-16.3, to 5.9) (Darrow et al., 2011).
Conversely, PM 2.5 nitrate was most negatively associated with birthweight in Whites, with a 10.3g (-21.0,
0.3) decrease observed, followed by Blacks 2.7g (-13.3, 7.9), and Hispanics 2.1g (-13.8, 9.4). Additionally,
PM 2.5 water-soluble metals also differed by race with Hispanics having the largest decrease in birthweight
of 18.7g (-32.2, -5.1), followed by Blacks 17.4g (29.8, -5.0), and Whites 14.7g (-27.1, -2.3) (Darrow et al.,
2011).
Combined, these findings highlight the pressing need of evaluating both PM 2.5 sources and their
respective components using the “gold standard” of personal exposure measurement to elicit whether it is
PM 2.5 mass, its sources (representing specific mixtures), or components that impact birthweight the most.
EXPOSURE MEASUREMENT ERROR OF SECONDHAND SMOKE IN PREGNANCY
STUDIES
Of these common PM 2.5 sources, secondhand smoke (SHS) is an endemic environmental exposure
that has been linked to several adverse pregnancy/birth outcomes including birthweight (Hawsawi et al.,
2015; Pogodina et al., 2009), preterm births (Hoyt et al., 2018), congenital malformations (Zheng et al.,
2019), and infant mortality (United States Surgeon General, 2014). However, because self-reported recall
is one of the most widely used methods to assess SHS exposure in health studies, and while many studies
have found good concordance between self-reported and objective measures (Kaufman et al., 2002; Kraev
et al., 2009; Vartiainen et al., 2002; Wipfli et al., 2008), misclassification of exposure is pervasive
(Cummings et al., 1990; Max et al., 2009; Woodruff et al., 2003). Exposure misclassification (or
measurement error) may be worse in pregnancy studies where social stigma and recall bias may impact

25
participants’ responses (Argalasova et al., 2019; Ashford et al., 2010; DeLorenze et al., 2002). This is
particularly worrisome considering that researchers found that just 17.9% (~35% in pregnancy studies) of
study questions were validated (Pérez-Ríos et al., 2013).
Objective measures for assessing exposure to SHS are available, and while they offer more precise
measurements of true exposure they generally also come with limitations, such as being logistically taxing,
costly, and invasive (Jaakkola & Jaakkola, 1997). Nicotine and its metabolite cotinine are common
objective assessments to measure exposure to SHS, and cotinine can be analyzed in blood, urine, saliva,
and hair matrices as a biomarker, with many considering this approach the “gold standard” for smoking and
SHS internal dose assessment (Ashford et al., 2010; Benowitz, 1996; M. M. Chen et al., 2021; DeLorenze
et al., 2002). Particle-based measurements in air samples offer an alternative to biospecimen-based
measurements and can take place through several formats, including stationary monitors placed indoors for
example (Kaur & Prasad, 2011; Semple et al., 2015; Williams et al., 1993), nicotine badges (Eisner et al.,
2001), and personal monitoring (Brook et al., 2011; Jenkins et al., 2001; Sloan et al., 2017). Validating self-
reported SHS exposure is a necessary step for health analyses in the absence of a standardized battery of
questions, especially for pregnancy studies, since the impact of measurement error in health analyses is
unknown, and where there is a lack of comparability of results across studies. In this dissertation, we
measured an optical property of PM 2.5 carbon components to estimate environmental tobacco smoke (ETS)
concentrations in personal PM 2.5 samples collected in the 3
rd
trimester (Lawless, 2004; Yan et al., 2011).
While objective measurements of SHS are preferred to reduce measurement error in exposure
assessment, the ease of use and cost-effectiveness, the ability to investigate longer duration exposures that
are often limited with objective measures due to logistical reasons, as well as the benefit of asking about
retrospective exposures that would be otherwise lost, make self-reported questionnaire methods a useful
tool. Therefore, there is a need to investigate the self-reported SHS questions in a health disparities
population and provide evidence for the use of another objective SHS measurement in a pregnancy study.

26
SUMMARY AND OVERALL AIMS
In summary, there seems to be a consistent association between outdoor PM 2.5 and lower
birthweight, which may be contributing to a host of negative health outcomes in both early and later life.
However, a great deal of heterogeneity in results exists, likely from regional or local differences in PM 2.5
composition, and measurement error of exposure from using ambient monitors to assign individual-level
exposure. The measurement error likely results in non-differential misclassification, which biases results
towards the null, potentially hiding the true magnitude and strength of the effect. There is a need to assess
the relationship between personal PM 2.5 and birthweight using “gold standard” personal exposure monitors
to potentially reduce the measurement error inherent with ambient air pollution research to explore this gap
in knowledge. Additionally, PM 2.5 is a mixture of hazardous substances from human-made and natural
sources, with different toxicities. Therefore, there is a need to assess different sources or origins of PM 2.5 to
determine which sources or components are the most detrimental, allowing more targeted
interventions/regulations. Additionally, due to the importance of SHS as a PM 2.5 source and endemic
environmental exposure, and its known adverse impacts on birth and other health outcomes during
pregnancy and later life, there is a need to evaluate typical approaches to SHS exposure assessment, namely,
self-report questionnaire data, by comparing them to an objective measure to investigate their ability to
capture “true” SHS exposure. Therefore, the primary aims of my dissertation are:
1) Investigate the effect of total personal PM 2.5 exposure on birthweight, while also evaluating the
impact of PM 2.5 origin or sources (roughly categorized as indoor vs outdoor) using
questionnaire data.
2) Estimate the relationship between single- and multi-pollutant sources and high-loading
components of personal PM 2.5 on birthweight in a health disparities population using a chemical
speciation and factor analysis approach from “gold standard” personal monitoring data.
Additionally, evaluate whether the association between major contributing personal PM 2.5
sources on birthweight differs by infant sex.

27
3) Describe measured airborne ETS in a personal exposure monitoring study of the MADRES
pregnancy cohort and evaluate the association between this objective measure and self-reported
SHS exposure to assess how well different questions employed in the MADRES cohort
adequately accounted for variation in measured airborne personal ETS exposure.

28

REFERENCES
Adams, K., Greenbaum, D. S., Shaikh, R., van Erp, A. M., & Russell, A. G. (2015). Particulate matter
components, sources, and health: Systematic approaches to testing effects. Journal of the Air &
Waste Management Association, 65(5), 544–558.
https://doi.org/10.1080/10962247.2014.1001884
Ar, B., Ak, P., C, P., R, D. M., A, F., N, A., Rd, L., Pp, S., & L, D. (2018). Low birth weight is associated
with increased fat intake in school-aged boys. The British Journal of Nutrition, 119(11).
https://doi.org/10.1017/S0007114518000892
Argalasova, L., Zitnanova, I., Vondrova, D., Dvorakova, M., Laubertova, L., Jurkovicova, J., Stofko, J.,
Weitzman, M., Waczulikova, I., & Simko, M. (2019). Self-Reported Exposure to ETS
(Environmental Tobacco Smoke), Urinary Cotinine, and Oxidative Stress Parameters in Pregnant
Women—The Pilot Study. International Journal of Environmental Research and Public Health,
16(9), 1656. https://doi.org/10.3390/ijerph16091656
Ashford, K. B., Hahn, E., Hall, L., Rayens, M. K., Noland, M., & Collins, R. (2010). Measuring prenatal
secondhand smoke exposure in mother–baby couplets. Nicotine & Tobacco Research, 12(2),
127–135. https://doi.org/10.1093/ntr/ntp185
Avery, C. L., Mills, K. T., Williams, R., McGraw, K. A., Poole, C., Smith, R. L., & Whitsel, E. A.
(2010). Estimating error in using ambient PM2.5 concentrations as proxies for personal
exposures. Epidemiology (Cambridge, Mass.), 21(2), 215–223.
https://doi.org/10.1097/EDE.0b013e3181cb41f7
Banerjee, T., Murari, V., Kumar, M., & Raju, M. P. (2015). Source apportionment of airborne particulates
through receptor modeling: Indian scenario. Atmospheric Research, 164–165, 167–187.
https://doi.org/10.1016/j.atmosres.2015.04.017
Banta, J. E., Ramadan, M., Alhusseini, N., Aloraini, K., & Modeste, N. (2021). Socio-demographics and
asthma prevalence, management, and outcomes among children 1–11 years of age in California.
Global Health Research and Policy, 6(1), 17. https://doi.org/10.1186/s41256-021-00199-y
Barker, D. J., & Osmond, C. (1986). Infant mortality, childhood nutrition, and ischaemic heart disease in
England and Wales. Lancet (London, England), 1(8489), 1077–1081.
https://doi.org/10.1016/s0140-6736(86)91340-1
Barker, D. J. P. (2004). The developmental origins of chronic adult disease. Acta Paediatrica (Oslo,
Norway: 1992). Supplement, 93(446), 26–33. https://doi.org/10.1111/j.1651-2227.2004.tb00236.x
Barnes, S. K., & Ozanne, S. E. (2011). Pathways linking the early environment to long-term health and
lifespan. Progress in Biophysics and Molecular Biology, 106(1), 323–336.
https://doi.org/10.1016/j.pbiomolbio.2010.12.005
Bastain, T. M., Chavez, T., Habre, R., Girguis, M. S., Grubbs, B., Toledo-Corral, C., Amadeus, M.,
Farzan, S. F., Al-Marayati, L., Lerner, D., Noya, D., Quimby, A., Twogood, S., Wilson, M.,
Chatzi, L., Cousineau, M., Berhane, K., Eckel, S. P., Lurmann, F., … Breton, C. (2019). Study
Design, Protocol and Profile of the Maternal And Developmental Risks from Environmental and
Social Stressors (MADRES) Pregnancy Cohort: A Prospective Cohort Study in Predominantly
Low-Income Hispanic Women in Urban Los Angeles. BMC Pregnancy and Childbirth, 19.
https://doi.org/10.1186/s12884-019-2330-7
Basu, R., Harris, M., Sie, L., Malig, B., Broadwin, R., & Green, R. (2014). Effects of fine particulate
matter and its constituents on low birth weight among full-term infants in California.
Environmental Research, 128, 42–51. https://doi.org/10.1016/j.envres.2013.10.008
Belbasis, L., Savvidou, M. D., Kanu, C., Evangelou, E., & Tzoulaki, I. (2016). Birth weight in relation to
health and disease in later life: An umbrella review of systematic reviews and meta-analyses.
BMC Medicine, 14(1), 147. https://doi.org/10.1186/s12916-016-0692-5

29
Bell, M. L., Belanger, K., Ebisu, K., Gent, J. F., Lee, H. J., Koutrakis, P., & Leaderer, B. P. (2010).
Prenatal Exposure to Fine Particulate Matter and Birth Weight. Epidemiology (Cambridge,
Mass.), 21(6), 884–891. https://doi.org/10.1097/EDE.0b013e3181f2f405
Bell, M. L., & Ebisu, K. (2012). Environmental Inequality in Exposures to Airborne Particulate Matter
Components in the United States. Environmental Health Perspectives, 120(12), 1699–1704.
https://doi.org/10.1289/ehp.1205201
Bell, M. L., Ebisu, K., & Peng, R. D. (2011). Community-level spatial heterogeneity of chemical
constituent levels of fine particulates and implications for epidemiological research. Journal of
Exposure Science & Environmental Epidemiology, 21(4), 372–384.
https://doi.org/10.1038/jes.2010.24
Benowitz, N. L. (1996). Cotinine as a Biomarker of Environmental Tobacco Smoke Exposure.
Epidemiologic Reviews, 18(2), 188–204. https://doi.org/10.1093/oxfordjournals.epirev.a017925
Boo, H. A. D., & Harding, J. E. (2006). The developmental origins of adult disease (Barker) hypothesis.
Australian and New Zealand Journal of Obstetrics and Gynaecology, 46(1), 4–14.
https://doi.org/10.1111/j.1479-828X.2006.00506.x
Bové, H., Bongaerts, E., Slenders, E., Bijnens, E. M., Saenen, N. D., Gyselaers, W., Van Eyken, P.,
Plusquin, M., Roeffaers, M. B. J., Ameloot, M., & Nawrot, T. S. (2019). Ambient black carbon
particles reach the fetal side of human placenta. Nature Communications, 10(1), 3866.
https://doi.org/10.1038/s41467-019-11654-3
Brook, R. D., Bard, R. L., Burnett, R. T., Shin, H. H., Vette, A., Croghan, C., Phillips, M., Rodes, C.,
Thornburg, J., & Williams, R. (2011). Differences in blood pressure and vascular responses
associated with ambient fine particulate matter exposures measured at the personal versus
community level. Occupational and Environmental Medicine, 68(3), 224–230.
https://doi.org/10.1136/oem.2009.053991
C, L., Y, H., D, H., L, L., F, X., G, L., C, G., F, L., S, C., C, W., & Y, Z. (2019). U-shaped relationship
between birth weight and childhood blood pressure in China. BMC Pediatrics, 19(1).
https://doi.org/10.1186/s12887-019-1638-9
Carroll, R. J. (2005). Measurement Error in Epidemiologic Studies. In Encyclopedia of Biostatistics.
American Cancer Society. https://doi.org/10.1002/0470011815.b2a03082
Caughey, R. W., & Michels, K. B. (2009). Birth weight and childhood leukemia: A meta-analysis and
review of the current evidence. International Journal of Cancer, 124(11), 2658–2670.
https://doi.org/10.1002/ijc.24225
CDC Newsroom. (2016, January 1). CDC. https://www.cdc.gov/media/releases/2017/p0718-diabetes-
report.html
Chen, C., Jin, Z., Yang, Y., Jiang, F., Huang, H., Liu, S., & Jin, X. (2019). Association of low birth
weight with thinness and severe obesity in children aged 3–12 years: A large-scale population-
based cross-sectional study in Shanghai, China. BMJ Open, 9(5).
https://doi.org/10.1136/bmjopen-2018-028738
Chen, M. M., Guo, S.-E., Yuan, C.-P., Okoli, C., & Liao, Y.-C. (2021). Association between Self-
Reported Survey Measures and Biomarkers of Second-Hand Tobacco Smoke Exposure in Non-
Smoking Pregnant Women. International Journal of Environmental Research and Public Health,
18(17), 9197. https://doi.org/10.3390/ijerph18179197
Clemente, D. B. P., Casas, M., Vilahur, N., Begiristain, H., Bustamante, M., Carsin, A.-E., Fernández, M.
F., Fierens, F., Gyselaers, W., Iñiguez, C., Janssen, B. G., Lefebvre, W., Llop, S., Olea, N.,
Pedersen, M., Pieters, N., Santa Marina, L., Souto, A., Tardón, A., … Nawrot, T. S. (2016).
Prenatal Ambient Air Pollution, Placental Mitochondrial DNA Content, and Birth Weight in the
INMA (Spain) and ENVIRONAGE (Belgium) Birth Cohorts. Environmental Health
Perspectives, 124(5), 659–665. https://doi.org/10.1289/ehp.1408981
Coker, E., Ghosh, J., Jerrett, M., Gomez-Rubio, V., Beckerman, B., Cockburn, M., Liverani, S., Su, J., Li,
A., Kile, M. L., Ritz, B., & Molitor, J. (2015). Modeling spatial effects of PM2.5 on term low

30
birth weight in Los Angeles County. Environmental Research, 142, 354–364.
https://doi.org/10.1016/j.envres.2015.06.044
Cummings, K. M., Markello, S. J., Mahoney, M., Bhargava, A. K., McElroy, P. D., & Marshall, J. R.
(1990). Measurement of current exposure to environmental tobacco smoke. Archives of
Environmental Health, 45(2), 74–79. https://doi.org/10.1080/00039896.1990.9935929
Dadvand, P., Ostro, B., Figueras, F., Foraster, M., Basagaña, X., Valentín, A., Martinez, D., Beelen, R.,
Cirach, M., Hoek, G., Jerrett, M., Brunekreef, B., & Nieuwenhuijsen, M. J. (2014). Residential
Proximity to Major Roads and Term Low Birth Weight: The Roles of Air Pollution, Heat, Noise,
and Road-Adjacent Trees. Epidemiology, 25(4), 518–525.
https://doi.org/10.1097/EDE.0000000000000107
Dadvand, P., Parker, J., Bell, M. L., Bonzini, M., Brauer, M., Darrow, L. A., Gehring, U., Glinianaia, S.
V., Gouveia, N., Ha, E., Leem, J. H., van den Hooven, E. H., Jalaludin, B., Jesdale, B. M.,
Lepeule, J., Morello-Frosch, R., Morgan, G. G., Pesatori, A. C., Pierik, F. H., … Woodruff, T. J.
(2013). Maternal Exposure to Particulate Air Pollution and Term Birth Weight: A Multi-Country
Evaluation of Effect and Heterogeneity. Environmental Health Perspectives, 121(3), 267–373.
https://doi.org/10.1289/ehp.1205575
Darrow, L. A., Klein, M., Strickland, M. J., Mulholland, J. A., & Tolbert, P. E. (2011). Ambient Air
Pollution and Birth Weight in Full-Term Infants in Atlanta, 1994–2004. Environmental Health
Perspectives, 119(5), 731–737. https://doi.org/10.1289/ehp.1002785
DeLorenze, G. N., Kharrazi, M., Kaufman, F. L., Eskenazi, B., & Bernert, J. T. (2002). Exposure to
environmental tobacco smoke in pregnant women: The association between self-report and serum
cotinine. Environmental Research, 90(1), 21–32. https://doi.org/10.1006/enrs.2001.4380
Di, Q., Amini, H., Shi, L., Kloog, I., Silvern, R., Kelly, J., Sabath, M. B., Choirat, C., Koutrakis, P.,
Lyapustin, A., Wang, Y., Mickley, L. J., & Schwartz, J. (2019). An ensemble-based model of
PM2.5 concentration across the contiguous United States with high spatiotemporal resolution.
Environment International, 130, 104909. https://doi.org/10.1016/j.envint.2019.104909
Ebisu, K., Belanger, K., & Bell, M. L. (2014). The Association between Airborne PM2.5 Chemical
Constituents and Birth Weight–Implication of Buffer Exposure Assignment. Environmental
Research Letters : ERL [Web Site], 9(8). https://doi.org/10.1088/1748-9326/9/8/084007
Eisner, M. D., Katz, P. P., Yelin, E. H., Hammond, S. K., & Blanc, P. D. (2001). Measurement of
environmental tobacco smoke exposure among adults with asthma. Environmental Health
Perspectives, 109(8), 809–814. https://doi.org/10.1289/ehp.01109809
Gee, S., Chin, D., Ackerson, L., Woo, D., & Howell, A. (2013). Prevalence of Childhood and Adolescent
Overweight and Obesity from 2003 to 2010 in an Integrated Health Care Delivery System.
Journal of Obesity, 2013. https://doi.org/10.1155/2013/417907
Giannadaki, D., Lelieveld, J., & Pozzer, A. (2016). Implementing the US air quality standard for PM2.5
worldwide can prevent millions of premature deaths per year. Environmental Health, 15(1), 88.
https://doi.org/10.1186/s12940-016-0170-8
Gray, S. C., Edwards, S. E., & Miranda, M. L. (2010). Assessing Exposure Metrics for PM and
Birthweight Models. Journal of Exposure Science & Environmental Epidemiology, 20(5), 469–
477. https://doi.org/10.1038/jes.2009.52
Gray, S. C., Edwards, S. E., Schultz, B. D., & Miranda, M. L. (2014). Assessing the impact of race, social
factors and air pollution on birth outcomes: A population-based study. Environmental Health, 13,
4. https://doi.org/10.1186/1476-069X-13-4
Habre, R., Coull, B., Moshier, E., Godbold, J., Grunin, A., Nath, A., Castro, W., Schachter, N., Rohr, A.,
Kattan, M., Spengler, J., & Koutrakis, P. (2014). Sources of indoor air pollution in New York
City residences of asthmatic children. Journal of Exposure Science & Environmental
Epidemiology, 24(3), 269–278. https://doi.org/10.1038/jes.2013.74
Hales, C. M. (2020). Prevalence of Obesity and Severe Obesity Among Adults: United States, 2017–2018.
360, 8.

31
Hales, C. N. (1997). Non-insulin-dependent diabetes mellitus. British Medical Bulletin, 53(1), 109–122.
https://doi.org/10.1093/oxfordjournals.bmb.a011594
Hales, C. N., & Barker, D. J. P. (2001). The thrifty phenotype hypothesis: Type 2 diabetes. British
Medical Bulletin, 60(1), 5–20. https://doi.org/10.1093/bmb/60.1.5
Harder, T., Rodekamp, E., Schellong, K., Dudenhausen, J. W., & Plagemann, A. (2007). Birth weight and
subsequent risk of type 2 diabetes: A meta-analysis. American Journal of Epidemiology, 165(8),
849–857. https://doi.org/10.1093/aje/kwk071
Harris, G., Thompson, W. D., Fitzgerald, E., & Wartenberg, D. (2014). The association of PM2.5 with
full term low birth weight at different spatial scales. Environmental Research, 134, 427–434.
https://doi.org/10.1016/j.envres.2014.05.034
Hawsawi, A. M., Bryant, L. O., & Goodfellow, L. T. (2015). Association Between Exposure to
Secondhand Smoke During Pregnancy and Low Birthweight: A Narrative Review. Respiratory
Care, 60(1), 135–140. https://doi.org/10.4187/respcare.02798
Hodan, W., & Barnard, W.(2004). Evaluating the Contribution of PM2.5 Precursor Gases and Re-
entrained Road Emissions to Mobile Source PM2.5 Particulate Matter Emissions. 58.
Hoffman, D. J., Reynolds, R. M., & Hardy, D. B. (2017). Developmental origins of health and disease:
Current knowledge and potential mechanisms. Nutrition Reviews, 75(12), 951–970.
https://doi.org/10.1093/nutrit/nux053
Hoyt, A. T., Canfield, M. A., Romitti, P. A., Botto, L. D., Anderka, M. T., Krikov, S. V., & Feldkamp, M.
L. (2018). Does Maternal Exposure to Secondhand Tobacco Smoke During Pregnancy Increase
the Risk for Preterm or Small-for-Gestational Age Birth? Maternal and Child Health Journal,
22(10), 1418–1429. https://doi.org/10.1007/s10995-018-2522-1
Huxley, R., Owen, C. G., Whincup, P. H., Cook, D. G., Rich-Edwards, J., Smith, G. D., & Collins, R.
(2007). Is birth weight a risk factor for ischemic heart disease in later life? The American Journal
of Clinical Nutrition, 85(5), 1244–1250. https://doi.org/10.1093/ajcn/85.5.1244
Isong, I. A., Rao, S. R., Bind, M.-A., Avendaño, M., Kawachi, I., & Richmond, T. K. (2018). Racial and
Ethnic Disparities in Early Childhood Obesity. Pediatrics, 141(1).
https://doi.org/10.1542/peds.2017-0865
Jaakkola, M. S., & Jaakkola, J. J. (1997). Assessment of exposure to environmental tobacco smoke.
European Respiratory Journal, 10(10), 2384–2397.
Jedrychowski, W., Bendkowska, I., Flak, E., Penar, A., Jacek, R., Kaim, I., Spengler, J. D., Camann, D.,
& Perera, F. P. (2004). Estimated Risk for Altered Fetal Growth Resulting from Exposure to Fine
Particles during Pregnancy: An Epidemiologic Prospective Cohort Study in Poland.
Environmental Health Perspectives, 112(14), 1398–1402. https://doi.org/10.1289/ehp.7065
Jedrychowski, W., Perera, F., Mrozek-Budzyn, D., Mroz, E., Flak, E., Spengler, J. D., Edwards, S., Jacek,
R., Kaim, I., & Skolicki, Z. (2009). Gender differences in fetal growth of newborns exposed
prenatally to airborne fine particulate matter. Environmental Research, 109(4), 447–456.
https://doi.org/10.1016/j.envres.2009.01.009
Jenkins, R. A., Maskarinec, M. P., Counts, R. W., Caton, J. E., Tomkins, B., & Ilgner, R. H. (2001).
Environmental tobacco smoke in an unrestricted smoking workplace: Area and personal exposure
monitoring. Journal of Exposure Science & Environmental Epidemiology, 11(5), 369–380.
https://doi.org/10.1038/sj.jea.7500178
Jornayvaz, F. R., Vollenweider, P., Bochud, M., Mooser, V., Waeber, G., & Marques-Vidal, P. (2016).
Low birth weight leads to obesity, diabetes and increased leptin levels in adults: The CoLaus
study. Cardiovascular Diabetology, 15(1), 73. https://doi.org/10.1186/s12933-016-0389-2
Kamai, E. M., McElrath, T. F., & Ferguson, K. K. (2019). Fetal growth in environmental epidemiology:
Mechanisms, limitations, and a review of associations with biomarkers of non-persistent chemical
exposures during pregnancy. Environmental Health, 18(1), 43. https://doi.org/10.1186/s12940-
019-0480-8
Karnes, J. H., Arora, A., Feng, J., Steiner, H. E., Sulieman, L., Boerwinkle, E., Clark, C., Cicek, M.,
Cohn, E., Gebo, K., Loperena-Cortes, R., Ohno-Machado, L., Mayo, K., Mockrin, S., Ramirez,

32
A., Schully, S., & Klimentidis, Y. C. (2021). Racial, ethnic, and gender differences in obesity and
body fat distribution: An All of Us Research Program demonstration project. PLOS ONE, 16(8),
e0255583. https://doi.org/10.1371/journal.pone.0255583
Kaufman, F. L., Kharrazi, M., Delorenze, G. N., Eskenazi, B., & Bernert, J. T. (2002). Estimation of
environmental tobacco smoke exposure during pregnancy using a single question on household
smokers versus serum cotinine. Journal of Exposure Science & Environmental Epidemiology,
12(4), 286–295. https://doi.org/10.1038/sj.jea.7500224
Kaur, J., & Prasad, V. M. (2011). Air Nicotine Monitoring for Second Hand Smoke Exposure in Public
Places in India. Indian Journal of Community Medicine : Official Publication of Indian
Association of Preventive & Social Medicine, 36(2), 98–103. https://doi.org/10.4103/0970-
0218.84126
Kioumourtzoglou, M.-A., Spiegelman, D., Szpiro, A. A., Sheppard, L., Kaufman, J. D., Yanosky, J. D.,
Williams, R., Laden, F., Hong, B., & Suh, H. (2014). Exposure measurement error in PM2.5
health effects studies: A pooled analysis of eight personal exposure validation studies.
Environmental Health, 13, 2. https://doi.org/10.1186/1476-069X-13-2
Kiserud, T., Benachi, A., Hecher, K., Perez, R. G., Carvalho, J., Piaggio, G., & Platt, L. D. (2018). The
World Health Organization fetal growth charts: Concept, findings, interpretation, and application.
American Journal of Obstetrics & Gynecology, 218(2), S619–S629.
https://doi.org/10.1016/j.ajog.2017.12.010
Korton, I., Ramsey, K., & Latzin, P. (2017). Air pollution during pregnancy and lung development in the
child—ClinicalKey. https://www-clinicalkey-com.libproxy1.usc.edu/#!/content/playContent/1-
s2.0-S1526054216300823?returnurl=null&referrer=null
Kraev, T., Adamkiewicz, G., Hammond, S., & Spengler, J. (2009). Indoor concentrations of nicotine in
low-income, multi-unit housing: Associations with smoking behaviours and housing
characteristics. Tobacco Control, 18(6), 438–444. https://doi.org/10.1136/tc.2009.029728
Lamichhane, D. K., Leem, J.-H., Lee, J.-Y., & Kim, H.-C. (2015). A meta-analysis of exposure to
particulate matter and adverse birth outcomes. Environmental Health and Toxicology, 30.
https://doi.org/10.5620/eht.e2015011
Laurent, O., Hu, J., Li, L., Cockburn, M., Escobedo, L., Kleeman, M. J., & Wu, J. (2014). Sources and
contents of air pollution affecting term low birth weight in Los Angeles County, California,
2001–2008. Environmental Research, 134, 488–495. https://doi.org/10.1016/j.envres.2014.05.003
Laurent, O., Hu, J., Li, L., Kleeman, M. J., Bartell, S. M., Cockburn, M., Escobedo, L., & Wu, J. (2016).
Low birth weight and air pollution in California: Which sources and components drive the risk?
Environment International, 92–93, 471–477. https://doi.org/10.1016/j.envint.2016.04.034
Lawless, P. (2004). Multiwavelength absorbance of filter deposits for determination of environmental
tobacco smoke and black carbon. Atmospheric Environment, 38(21), 3373–3383.
https://doi.org/10.1016/j.atmosenv.2004.03.038
Li, Y., Lu, Q., Wang, Y., & Ma, S. (2020). Racial differences in testicular cancer in the United States:
Descriptive epidemiology. BMC Cancer, 20(1), 284. https://doi.org/10.1186/s12885-020-06789-2
Li, Z., Li, X., Song, H., Tao, B., Qiu, B., Tian, D., Zhan, M., Wu, Z., Wu, J., Zhang, Q., & Wang, J.
(2021). Effects of short-term ambient PM2.5 exposure on the blood cell count and hemoglobin
concentration among 82,431 people in eastern China. Science of The Total Environment, 776,
146046. https://doi.org/10.1016/j.scitotenv.2021.146046
Li, Z., Tang, Y., Song, X., Lazar, L., Li, Z., & Zhao, J. (2019). Impact of ambient PM2.5 on adverse birth
outcome and potential molecular mechanism. Ecotoxicology and Environmental Safety, 169,
248–254. https://doi.org/10.1016/j.ecoenv.2018.10.109
Lin, C.-C., & Santolaya-Forgas, J. (1998). CURRENT CONCEPTS OF FETAL GROWTH
RESTRICTION: PART I. CAUSES, CLASSIFICATION, AND PATHOPHYSIOLOGY.
[Review]. Obstetrics & Gynecology, 92(6), 1044–1055.

33
M, Z., F, S., E, R., Fn, N., H, K., Mr, H., & Z, A. (2013). Does low birth weight predict hypertension and
obesity in schoolchildren? Annals of Nutrition & Metabolism, 63(1–2).
https://doi.org/10.1159/000351869
Madsen, C., Gehring, U., Erik Walker, S., Brunekreef, B., Stigum, H., Næss, Ø., & Nafstad, P. (2010).
Ambient air pollution exposure, residential mobility and term birth weight in Oslo, Norway.
Environmental Research, 110(4), 363–371. https://doi.org/10.1016/j.envres.2010.02.005
Martin, J. A., Hamilton, B. E., Osterman, M. J. K., & Driscoll, A. K. (2018). National Vital Statistics
Reports Volume 68, Number 13, November 30, 2019, Births: Final Data for 2018. 47.
Max, W., Sung, H.-Y., & Shi, Y. (2009). Who Is Exposed to Secondhand Smoke? Self-Reported and
Serum Cotinine Measured Exposure in the U.S., 1999–2006. International Journal of
Environmental Research and Public Health, 6(5), 1633–1648.
https://doi.org/10.3390/ijerph6051633
Morrison, J. L., Duffield, J. A., Muhlhausler, B. S., Gentili, S., & McMillen, I. C. (2010). Fetal growth
restriction, catch-up growth and the early origins of insulin resistance and visceral obesity.
Pediatric Nephrology (Berlin, Germany), 25(4), 669–677. https://doi.org/10.1007/s00467-009-
1407-3
Nachman, R. M., Mao, G., Zhang, X., Hong, X., Chen, Z., Soria, C. S., He, H., Wang, G., Caruso, D.,
Pearson, C., Biswal, S., Zuckerman, B., Wills-Karp, M., & Wang, X. (2016). Intrauterine
Inflammation and Maternal Exposure to Ambient PM2.5 during Preconception and Specific
Periods of Pregnancy: The Boston Birth Cohort. Environmental Health Perspectives, 124(10),
1608–1615. https://doi.org/10.1289/EHP243
Nardozza, L. M. M., Caetano, A. C. R., Zamarian, A. C. P., Mazzola, J. B., Silva, C. P., Marçal, V. M. G.,
Lobo, T. F., Peixoto, A. B., & Araujo Júnior, E. (2017). Fetal growth restriction: Current
knowledge. Archives of Gynecology and Obstetrics, 295(5), 1061–1077.
https://doi.org/10.1007/s00404-017-4341-9
National Academies of Sciences, E. (2017). Communities in Action: Pathways to Health Equity.
https://doi.org/10.17226/24624
Ng, C., Malig, B., Hasheminassab, S., Sioutas, C., Basu, R., & Ebisu, K. (2017). Source apportionment of
fine particulate matter and risk of term low birth weight in California: Exploring modification by
region and maternal characteristics. Science of The Total Environment, 605–606, 647–654.
https://doi.org/10.1016/j.scitotenv.2017.06.053
Oksuzyan, S., Crespi, C., Cockburn, M., Mezei, G., Vergara, X., & Kheifets, L. (2015). Race/ethnicity
and the risk childhood leukemia: A case-control study in California. Journal of Epidemiology and
Community Health, 69(8), 795–802. https://doi.org/10.1136/jech-2014-204975
Pedersen, M., Giorgis-Allemand, L., Bernard, C., Aguilera, I., Andersen, A.-M. N., Ballester, F., Beelen,
R. M. J., Chatzi, L., Cirach, M., Danileviciute, A., Dedele, A., Eijsden, M. van, Estarlich, M.,
Fernández-Somoano, A., Fernández, M. F., Forastiere, F., Gehring, U., Grazuleviciene, R.,
Gruzieva, O., … Slama, R. (2013). Ambient air pollution and low birthweight: A European
cohort study (ESCAPE). The Lancet Respiratory Medicine, 1(9), 695–704.
https://doi.org/10.1016/S2213-2600(13)70192-9
Pérez-Ríos, M., Schiaffino, A., López, M. J., Nebot, M., Galán, I., Fu, M., Martínez-Sánchez, J. M.,
Moncada, A., Montes, A., Ariza, C., & Fernández, E. (2013). Questionnaire-based second-hand
smoke assessment in adults. European Journal of Public Health, 23(5), 763–767.
https://doi.org/10.1093/eurpub/cks069
Pogodina, C., Brunner Huber, L. R., Racine, E. F., & Platonova, E. (2009). Smoke-Free Homes for
Smoke-Free Babies: The Role of Residential Environmental Tobacco Smoke on Low Birth
Weight. Journal of Community Health, 34(5), 376–382. https://doi.org/10.1007/s10900-009-
9169-1
Priante, E., Verlato, G., Giordano, G., Stocchero, M., Visentin, S., Mardegan, V., & Baraldi, E. (2019).
Intrauterine Growth Restriction: New Insight from the Metabolomic Approach. Metabolites,
9(11), 267. https://doi.org/10.3390/metabo9110267

34
Qiao, Y., Ma, J., Wang, Y., Li, W., Katzmarzyk, P. T., Chaput, J.-P., Fogelholm, M., Johnson, W. D.,
Kuriyan, R., Kurpad, A., Lambert, E. V., Maher, C., Maia, J., Matsudo, V., Olds, T., Onywera,
V., Sarmiento, O. L., Standage, M., Tremblay, M. S., … ISCOLE Research Group. (2015). Birth
weight and childhood obesity: A 12-country study. International Journal of Obesity Supplements,
5(Suppl 2), S74-79. https://doi.org/10.1038/ijosup.2015.23
Richmond-Bryant, J., & Long, T. C. (2020). Influence of exposure measurement errors on results from
epidemiologic studies of different designs. Journal of Exposure Science & Environmental
Epidemiology, 30(3), 420–429. https://doi.org/10.1038/s41370-019-0164-z
Risnes, K. R., Vatten, L. J., Baker, J. L., Jameson, K., Sovio, U., Kajantie, E., Osler, M., Morley, R.,
Jokela, M., Painter, R. C., Sundh, V., Jacobsen, G. W., Eriksson, J. G., Sørensen, T. I. A., &
Bracken, M. B. (2011). Birthweight and mortality in adulthood: A systematic review and meta-
analysis. International Journal of Epidemiology, 40(3), 647–661.
https://doi.org/10.1093/ije/dyq267
Rossen, L. M. (2014). Neighbourhood economic deprivation explains racial/ethnic disparities in
overweight and obesity among children and adolescents in the USA. Journal of Epidemiology
and Community Health, 68(2), 123–129. https://doi.org/10.1136/jech-2012-202245
Savitz, D. A., Bobb, J. F., Carr, J. L., Clougherty, J. E., Dominici, F., Elston, B., Ito, K., Ross, Z., Yee,
M., & Matte, T. D. (2014). Ambient Fine Particulate Matter, Nitrogen Dioxide, and Term Birth
Weight in New York, New York. American Journal of Epidemiology, 179(4), 457–466.
https://doi.org/10.1093/aje/kwt268
Schembari, A., de Hoogh, K., Pedersen, M., Dadvand, P., Martinez, D., Hoek, G., Petherick, E. S.,
Wright, J., & Nieuwenhuijsen, M. J. (2015). Ambient Air Pollution and Newborn Size and
Adiposity at Birth: Differences by Maternal Ethnicity (the Born in Bradford Study Cohort).
Environmental Health Perspectives, 123(11), 1208–1215. https://doi.org/10.1289/ehp.1408675
Semple, S., Apsley, A., Azmina Ibrahim, T., Turner, S. W., & Cherrie, J. W. (2015). Fine particulate
matter concentrations in smoking households: Just how much secondhand smoke do you breathe
in if you live with a smoker who smokes indoors? Tobacco Control, 24(e3), e205-211.
https://doi.org/10.1136/tobaccocontrol-2014-051635
Sloan, C. D., Weber, F. X., Bradshaw, R. K., Philipp, T. J., Barber, W. B., Palmer, V. L., Graul, R. J.,
Tuttle, S. C., Chartier, R. T., & Johnston, J. D. (2017). Elemental analysis of infant airborne
particulate exposures. Journal of Exposure Science & Environmental Epidemiology, 27(5), 526–
534. https://doi.org/10.1038/jes.2016.77
Smith, C. J., Ryckman, K. K., Barnabei, V. M., Howard, B. V., Isasi, C. R., Sarto, G. E., Tom, S. E., Van
Horn, L. V., Wallace, R. B., & Robinson, J. G. (2016). The impact of birth weight on
cardiovascular disease risk in the Women’s Health Initiative. Nutrition, Metabolism, and
Cardiovascular Diseases: NMCD, 26(3), 239–245. https://doi.org/10.1016/j.numecd.2015.10.015
Smith, L., Luchini, C., Demurtas, J., Soysal, P., Stubbs, B., Hamer, M., Nottegar, A., Lawlor, R. T.,
Lopez-Sanchez, G. F., Firth, J., Koyanagi, A., Roberts, J., Willeit, P., Waldhoer, T., Loosemore,
M., Abbs, A. D., Johnstone, J., Yang, L., & Veronese, N. (2019). Telomere length and health
outcomes: An umbrella review of systematic reviews and meta-analyses of observational studies.
Ageing Research Reviews, 51, 1–10. https://doi.org/10.1016/j.arr.2019.02.003
Sørensen, M., Daneshvar, B., Hansen, M., Dragsted, L. O., Hertel, O., Knudsen, L., & Loft, S. (2003).
Personal PM2.5 exposure and markers of oxidative stress in blood. Environmental Health
Perspectives, 111(2), 161–166. https://doi.org/10.1289/ehp.111-1241344
Spanakis, E. K., & Golden, S. H. (2013). Race/Ethnic Difference in Diabetes and Diabetic Complications.
Current Diabetes Reports, 13(6), 10.1007/s11892-013-0421–0429.
https://doi.org/10.1007/s11892-013-0421-9
Spracklen, C. N., Wallace, R. B., Sealy-Jefferson, S., Robinson, J. G., Freudenheim, J. L., Wellons, M. F.,
Saftlas, A. F., Snetselaar, L. G., Manson, J. E., Hou, L., Qi, L., Chlebowski, R. T., & Ryckman,
K. K. (2014). Birth Weight and Subsequent Risk of Cancer. Cancer Epidemiology, 38(5), 538–
543. https://doi.org/10.1016/j.canep.2014.07.004

35
Stansfield, B. K., Fain, M. E., Bhatia, J., Gutin, B., Nguyen, J. T., & Pollock, N. K. (2016). Nonlinear
Relationship between Birth Weight and Visceral Fat in Adolescents. The Journal of Pediatrics,
174, 185–192. https://doi.org/10.1016/j.jpeds.2016.04.012
Stapleton, P. A. (2016). Gestational nanomaterial exposures: Microvascular implications during
pregnancy, fetal development and adulthood. The Journal of Physiology, 594(8), 2161–2173.
https://doi.org/10.1113/JP270581
Stieb, D. M., Chen, L., Eshoul, M., & Judek, S. (2012). Ambient air pollution, birth weight and preterm
birth: A systematic review and meta-analysis. Environmental Research, 117, 100–111.
https://doi.org/10.1016/j.envres.2012.05.007
Sun, X., Luo, X., Zhao, C., Zhang, B., Tao, J., Yang, Z., Ma, W., & Liu, T. (2016). The associations
between birth weight and exposure to fine particulate matter (PM2.5) and its chemical
constituents during pregnancy: A meta-analysis. Environmental Pollution, 211, 38–47.
https://doi.org/10.1016/j.envpol.2015.12.022
Tessum, C. W., Paolella, D. A., Chambliss, S. E., Apte, J. S., Hill, J. D., & Marshall, J. D. (2021). PM2.5
polluters disproportionately and systemically affect people of color in the United States. Science
Advances, 7(18), eabf4491. https://doi.org/10.1126/sciadv.abf4491
Umer, A., Hamilton, C., Cottrell, L., Giacobbi, P., Innes, K., Kelley, G. A., Neal, W., John, C., & Lilly,
C. (2020). Association between birth weight and childhood cardiovascular disease risk factors in
West Virginia. Journal of Developmental Origins of Health and Disease, 11(1), 86–95.
https://doi.org/10.1017/S204017441900045X
United States Surgeon General. (2014). The Health Consequences of Smoking -- 50 Years of progress: A
Report of the Surgeon General: (510072014-001) [Data set]. American Psychological
Association. https://doi.org/10.1037/e510072014-001
Upadhyay, R. P., Naik, G., Choudhary, T. S., Chowdhury, R., Taneja, S., Bhandari, N., Martines, J. C.,
Bahl, R., & Bhan, M. K. (2019). Cognitive and motor outcomes in children born low birth
weight: A systematic review and meta-analysis of studies from South Asia. BMC Pediatrics,
19(1), 35. https://doi.org/10.1186/s12887-019-1408-8
US EPA, O. (2014, April 10). NAAQS Table [Other Policies and Guidance]. https://www.epa.gov/criteria-
air-pollutants/naaqs-table
US EPA, O. (2016, April 19). Particulate Matter (PM) Basics [Overviews and Factsheets].
https://www.epa.gov/pm-pollution/particulate-matter-pm-basics
Vaag, A. A., Grunnet, L. G., Arora, G. P., & Brøns, C. (2012). The thrifty phenotype hypothesis revisited.
Diabetologia, 55(8), 2085–2088. https://doi.org/10.1007/s00125-012-2589-y
Vartiainen, E., Seppälä, T., Lillsunde, P., & Puska, P. (2002). Validation of self reported smoking by
serum cotinine measurement in a community-based study. Journal of Epidemiology &
Community Health, 56(3), 167–170. https://doi.org/10.1136/jech.56.3.167
Vilanova, C. S., Hirakata, V. N., de Souza Buriol, V. C., Nunes, M., Goldani, M. Z., & da Silva, C. H.
(2019). The relationship between the different low birth weight strata of newborns with infant
mortality and the influence of the main health determinants in the extreme south of Brazil.
Population Health Metrics, 17(1), 15. https://doi.org/10.1186/s12963-019-0195-7
Visentin, S., Grumolato, F., Nardelli, G. B., Di Camillo, B., Grisan, E., & Cosmi, E. (2014). Early origins
of adult disease: Low birth weight and vascular remodeling. Atherosclerosis, 237(2), 391–399.
https://doi.org/10.1016/j.atherosclerosis.2014.09.027
Wallace, L. A., Mitchell, H., O'Connor, G. T., Neas, L., Lippmann, M., Kattan, M., Koenig, J., Stout, J.
W., Vaughn, B. J., Wallace, D., Walter, M., Adams, K., Liu, L. J., & Inner-City Asthma Study
(2003). Particle concentrations in inner-city homes of children with asthma: the effect of
smoking, cooking, and outdoor pollution. Environmental health perspectives, 111(9), 1265–1272.
https://doi.org/10.1289/ehp.6135
Wang, L., Guo, P., Tong, H., Wang, A., Chang, Y., Guo, X., Gong, J., Song, C., Wu, L., Wang, T.,
Hopke, P. K., Chen, X., Tang, N., & Mao, H. (2020). Traffic-related metrics and adverse birth

36
outcomes: A systematic review and meta-analysis. Environmental Research, 188, 109752.
https://doi.org/10.1016/j.envres.2020.109752
Watkins, W. J., Kotecha, S. J., & Kotecha, S. (2016). All-Cause Mortality of Low Birthweight Infants in
Infancy, Childhood, and Adolescence: Population Study of England and Wales. PLoS Medicine,
13(5). https://doi.org/10.1371/journal.pmed.1002018
Wei, H., Zhang, C., & Silveyra, P. (2020). The Relationships Between Prenatal Smoking Exposure and
Telomere Lengths in Fetuses, Infants, and Children: A Systematic Literature Review. Journal of
Addictions Nursing, 31(4), 243–252. https://doi.org/10.1097/JAN.0000000000000364
Wentz, A. E., Messer, L. C., Nguyen, T., & Boone-Heinonen, J. (2014). Small and large size for
gestational age and neighborhood deprivation measured within increasing proximity to homes.
Health & Place, 30, 98–106. https://doi.org/10.1016/j.healthplace.2014.08.010
Whitaker, A. H., Feldman, J. F., Lorenz, J. M., Shen, S., McNicholas, F., Nieto, M., McCulloch, D.,
Pinto-Martin, J. A., & Paneth, N. (2006). Motor and Cognitive Outcomes in Nondisabled Low-
Birth-Weight Adolescents: Early Determinants. Archives of Pediatrics & Adolescent Medicine,
160(10), 1040. https://doi.org/10.1001/archpedi.160.10.1040
Wilhelm, M., Ghosh, J. K., Su, J., Cockburn, M., Jerrett, M., & Ritz, B. (2012). Traffic-Related Air
Toxics and Term Low Birth Weight in Los Angeles County, California. Environmental Health
Perspectives, 120(1), 132–138. https://doi.org/10.1289/ehp.1103408
Williams, R., Collier, A., & Lewtas, J. (1993). Environmental Tobacco Smoke Exposure of Young
Children as Assessed Using a Passive Diffusion Device for Nicotine. Indoor Environment, 2, 98–
104. https://doi.org/10.1159/000463231
Wipfli, H., Avila-Tang, E., Navas-Acien, A., Kim, S., Onicescu, G., Yuan, J., Breysse, P., & Samet, J. M.
(2008). Secondhand Smoke Exposure Among Women and Children: Evidence From 31
Countries. American Journal of Public Health, 98(4), 672–679.
https://doi.org/10.2105/AJPH.2007.126631
Wojtyla, C., Zielinska, K., Wojtyla-Buciora, P., & Panek, G. (2020). Prenatal Fine Particulate Matter
(PM2.5) Exposure and Pregnancy Outcomes—Analysis of Term Pregnancies in Poland.
International Journal of Environmental Research and Public Health, 17(16).
https://doi.org/10.3390/ijerph17165820
Womack, L. S., Rossen, L. M., & Martin, J. A. (2018, March 1). Singleton Low Birthweight Rates, by
Race and Hispanic Origin: United States, 2006–2016 [Center for Disease Control and Prevention
(CDC)]. National Center for Health Statistics.
https://www.cdc.gov/nchs/products/databriefs/db306.htm
Woodruff, S. I., Conway, T. L., Edwards, C. C., & Hovell, M. F. (2003). Acceptability and validity of
hair collection from Latino children to assess exposure to environmental tobacco smoke. Nicotine
& Tobacco Research, 5(3), 375–385. https://doi.org/10.1080/1462220031000094187
Yan, B., Kennedy, D., Miller, R. L., Cowin, J. P., Jung, K., Perzanowski, M., Balletta, M., Perera, F. P.,
Kinney, P. L., & Chillrud, S. N. (2011). Validating a nondestructive optical method for
apportioning colored particulate matter into black carbon and additional components.
Atmospheric Environment (Oxford, England : 1994), 45(39), 7478–7486.
https://doi.org/10.1016/j.atmosenv.2011.01.044
Yuan, L., Zhang, Y., Wang, W., Chen, R., Liu, Y., Liu, C., Kan, H., Gao, Y., & Tian, Y. (2020). Critical
windows for maternal fine particulate matter exposure and adverse birth outcomes: The Shanghai
birth cohort study. Chemosphere, 240, 124904.
https://doi.org/10.1016/j.chemosphere.2019.124904
Zeger, S. L., Thomas, D., Dominici, F., Samet, J. M., Schwartz, J., Dockery, D., & Cohen, A. (2000).
Exposure measurement error in time-series studies of air pollution: Concepts and consequences.
Environmental Health Perspectives, 108(5), 419–426.
Zheng, Z., Xie, G., Yang, T., & Qin, J. (2019). Congenital malformations are associated with secondhand
smoke among nonsmoking women: A meta-analysis. Birth (Berkeley, Calif.), 46(2), 222–233.
https://doi.org/10.1111/birt.12401

37
Zhu, X., Liu, Y., Chen, Y., Yao, C., Che, Z., & Cao, J. (2015). Maternal exposure to fine particulate
matter (PM2.5) and pregnancy outcomes: A meta-analysis. Environmental Science and Pollution
Research, 22(5), 3383–3396. https://doi.org/10.1007/s11356-014-3458-7
Zhu, Y., Sidell, M. A., Arterburn, D., Daley, M. F., Desai, J., Fitzpatrick, S. L., Horberg, M. A.,
Koebnick, C., McCormick, E., Oshiro, C., Young, D. R., & Ferrara, A. (2019). Racial/Ethnic
Disparities in the Prevalence of Diabetes and Prediabetes by BMI: Patient Outcomes Research To
Advance Learning (PORTAL) Multisite Cohort of Adults in the U.S. Diabetes Care, 42(12),
2211–2219. https://doi.org/10.2337/dc19-0532

38
CHAPTER 2
IN-UTERO PERSONAL EXPOSURE TO PM 2.5 IMPACTED BY INDOOR AND OUTDOOR
SOURCES IN THE MADRES COHORT
ABSTRACT
Background: In-utero exposure to outdoor particulate matter with aerodynamic diameter less than 2.5µm
(PM 2.5) is linked with low birthweight. However, previous results are mixed, likely due to measurement
error introduced by estimating personal exposure from ambient data. We investigated the effect of total
personal PM 2.5 exposure on birthweight and whether it differed when it was more heavily impacted by
sources of indoor vs outdoor origin in the MADRES cohort study.
Methods: Personal PM 2.5 exposure was measured in 205 pregnant women in the 3
rd
trimester using 48-hour
integrated, filter-based sampling. Linear regression was used to test the association between personal PM 2.5
exposure and birthweight, adjusting for key covariates. Interactions of PM 2.5 with variables representing
indoor sources of PM 2.5, home ventilation, or time spent indoors tested whether the effect of total PM 2.5 on
birthweight varied when it was more impacted by sources of indoor vs outdoor origin.
Results: In a sample of largely Hispanic (81%) pregnant women, total personal PM 2.5 was not significantly
associated with birthweight (β=38.6 per 1SD increase in PM2.5; 95% CI:-21.1, 98.2). This association
however, differed by home type (single family home: 156.9 (26.9, 287.0), 2-4 attached units:-16.6 (-111.9,
78.7), 5+ units:-62.6 (-184.9, 59.6), missing: 145.4 (-4.1, 294.9), interaction p=0.028) and by household air
conditioner use (none of the time: -27.6 (-101.5, 46.3) vs. some of the time: 139.9 (42.9, 237.0), interaction
p=0.008). Additionally, the effect of personal PM 2.5 on birthweight varied by time spent indoors (none or
little of the time:-45.1 (-208.3, 118.1) vs. most or all of the time: 57.1 (-7.3, 121.6), interaction p=0.255).
Conclusions: While we found no significant association between total personal PM 2.5 exposure and
birthweight, there was evidence that multi-unit housing (vs. single-family homes), candle and/or incense
smoke, and greater outdoor source contributions to personal PM 2.5 were more strongly associated with lower
birthweight.

39
INTRODUCTION
In the United States (U.S.), an estimated 8.3% of newborns are born with low birthweight (LBW)
(Martin et al., 2018); defined as below 2,500 grams (g). The impact of LBW is far reaching, with research
showing it is associated with infant mortality (Vilanova et al., 2019; Watkins et al., 2016) and later life
obesity (Jornayvaz et al., 2016), type-2 diabetes (Mi et al., 2017), cardiovascular disease (Huxley et al.,
2007; Risnes et al., 2011; Smith et al., 2016; Umer et al., 2020), and impaired cognitive development
(Upadhyay et al., 2019; Whitaker et al., 2006). Within the U.S., such health outcomes are often
disproportionate with regard to race/ethnicity, with obesity and type-2 diabetes prevalence highest in
Hispanic and Black populations across the lifetime (Petersen, 2019; Rossen, 2014).
In the past decade, several epidemiological studies have established a relationship between outdoor
air pollution exposure and birthweight and/or LBW (Lamichhane et al., 2015; Li et al., 2017; Pedersen et
al., 2013; Stieb et al., 2012). Within the US, these studies primarily focused on federally-regulated criteria
air pollutants. One of those is particulate matter (PM) with an aerodynamic diameter less than 2.5µm (PM 2.5)
(Dadvand et al., 2014; Huang et al., 2015; Rich et al., 2015; Savitz et al., 2014; Schembari et al., 2015).

In-
utero PM 2.5 exposure is hypothesized to create a hostile intrauterine environment likely resulting from
oxidative stress, DNA methylation changes, mitochondrial DNA content alteration, and endocrine
disruption (Clemente et al., 2016; Li et al., 2019). Such mechanistic alterations may lead to health risks in
later life such as the development of visceral adiposity and altered glucose homeostasis (Barnes and
Ozanne, 2011; Hales, 1997; Morrison et al., 2010; Visentin et al., 2014).
While several reviews have concluded a weak to moderate association between outdoor PM 2.5 and
several birth outcomes, including a decrease in birthweight and an increased risk of LBW (Li et al., 2017;
Stieb et al., 2016; Sun et al., 2016), the literature remains inconsistent. Reductions in birthweight due to
outdoor PM 2.5 exposure also vary by race/ethnicity (Basu et al., 2014), possibly due to Hispanic and Black
mothers experiencing the greatest burden of air pollution exposure (Bell and Ebisu, 2012; Mikati et al.,
2018). Effect estimates also differ depending on the exposure window under study, with the 3
rd
trimester
showing the most consistent evidence of greater risk of LBW (Dadvand et al., 2014; Savitz et al., 2014;

40
Schembari et al., 2015; Zhu et al., 2015). Additionally, most health studies to date estimated an individual’s
exposure to outdoor PM 2.5 at the residential level using models that typically incorporate ambient
monitoring, remote sensing, and/or geospatial data (Dadvand et al., 2013; Ebisu et al., 2014; Gray et al.,
2010; Harris et al., 2014). While these models are increasingly capable of capturing spatial variability in
outdoor air pollution, they inherently suffer from exposure measurement error in terms of estimating
personal exposure to air pollution of outdoor origin, which might bias effect estimates and attenuate power
to detect health effects (Carroll, 2005; Kioumourtzoglou et al., 2014; Zeger et al., 2000). This is because
individuals spend the majority of their time indoors, and their “true” personal exposure to PM 2.5 of outdoor
origin is a result of the infiltration efficiency of PM 2.5 indoors and time-activity patterns, most accurately
captured by personal monitoring (Gray et al., 2011). Finally, there is currently very little research into the
effect of total personal exposure to PM 2.5 prenatally on birthweight. Total personal PM 2.5 is impacted by
multiple sources including personal activity, indoor sources, and outdoor sources (or PM 2.5 of outdoor
origin, which may only represent a small fraction of an individual’s total personal PM 2.5 (Habre et al., 2014).
Therefore, quantifying the influence of total personal PM 2.5 on birthweight is also an important question
that has not yet been thoroughly investigated.
Personal monitoring of air pollution is a sophisticated, yet often expensive and burdensome method
of exposure assessment, and as such, only a small number of studies have used it, the majority of which
have focused on toxic polyaromatic hydrocarbons (PAHs) (Choi et al., 2012, 2008; Rundle et al., 2012).
One study found an inverse association between personal PM 2.5 and birthweight (Jedrychowski et al., 2009).
However, to our knowledge, very few studies have been conducted in a health disparities population with
potentially greater exposure to PM 2.5 of outdoor origin and greater vulnerability or susceptibility to its
effects (Morello-Frosch et al., 2011), particularly in the 3
rd
trimester where most fetal weight gain occurs
(Kiserud et al., 2018). Additionally, there is a pressing need to evaluate the effects of PM 2.5 impacted by
sources of indoor vs outdoor origin (hereinafter referred to as indoor vs outdoor sources for simplicity) due
to the differences in their chemical composition and thus potential toxicity, and the fact that only ambient
PM 2.5 concentrations are regulated .

41
Therefore, the purpose of this present study was to bridge these gaps in knowledge by evaluating
the role of 3
rd
trimester personal PM 2.5 exposure on birth weight in a health disparities population in Los
Angeles, CA. In addition, we investigated whether the effect of total personal PM 2.5 on birthweight was
different when it was more impacted by indoor vs outdoor sources. To accomplish this, we tested
interactions with questionnaire-based variables that correlate directly with greater indoor (e.g., indoor
burning of candles or incense) or outdoor (e.g., time spent outdoors) contributions to total personal PM 2.5.
MATERIALS AND METHODS
Study Population
The Maternal and Developmental Risks from Environmental and Social Stressors (MADRES)
study is an ongoing prospective cohort study of ~900 pregnant, primarily Hispanic, low-income mothers in
Los Angeles County, motivated to investigate the cumulative impact of environmental pollutants and
psychosocial, behavioral, and built environmental risk factors on maternal and infant health outcomes
(Bastain et al., 2019). Pregnant women were enrolled via partnerships with four prenatal care providers in
Los Angeles beginning November 2015, including one county hospital clinic, two non-profit community
health clinics, and a private obstetrics and gynecology practice.
Participant eligibility included: 1) at least 18 years old, 2) fluency in either Spanish or English, and
3) less than 30-weeks gestation at recruitment. Exclusion criteria for the study included: 1) multiple
gestation, 2) current incarceration, 3) HIV positive, and 4) having a physical, mental, or cognitive disability
that would prevent the participant from providing informed consent.
The current analysis leverages data collected as part of a 214-participant personal PM 2.5 exposure
monitoring study nested within the MADRES cohort. Women were asked to wear a crossbody sampling
purse with a personal monitoring apparatus for a 48-hour monitoring period in the 3
rd
trimester. This subset
was comparable to the larger MADRES cohort in terms of demographics, birth outcomes, and outdoor air
pollution metrics.

42
Personal PM 2.5 Exposure Monitoring
Total, 48-hour integrated personal PM 2.5 exposure was measured in the 3
rd
trimester using a custom
sampling design on a subset of 214 women recruited from the larger cohort between October 2016 and
February 2020. A trained, bilingual study staff member recruited participants during one of their 3
rd

trimester study visits at the University of Southern California (USC) clinic. Participants were provided with
a personal sampling crossbody purse containing a Gilian Plus Datalogging Pump (Sensidyne Inc.,
Clearwater, FL), which was programmed to start at midnight (the following day) and actively sample at a
50% cycle and flow rate of 1.8 liters per minute (LPM). The pump was connected to a Harvard PM 2.5
personal environmental monitor (PEM) with a 37mm Pall Teflo filter. Staff members provided instructions
regarding proper use and demonstrated how to wear the sampling bag, with the sampling inlet located on
the purse strap in the shoulder area around the breathing zone.
Participants were instructed to wear the sampling device during all waking hours while going about
their normal daily activities. Exceptions to this requirement included while performing potentially
dangerous activities (e.g. driving), showering, sleeping, or otherwise unable to. Participants were asked to
protect the sampling device from water, high humidity (such as showering or sauna), heat, pets, and
children. When they could not wear the monitor continuously, such as when sleeping or driving, they were
asked to place it on a bedside table or beside them on the passenger seat, away from surfaces as much as
possible and unobstructed. Additionally, when not wearing the monitor, individuals were asked to keep the
monitor elevated from the ground and away from the walls due to sampling artifacts that could result from
resuspended dust or removal on surfaces, respectively.
The sampling pump was programmed to shut down after the 48-hour sampling period was
completed, and study staff coordinated device pickup and conducted a brief exit survey with participants
the following day. When the sampling devices arrived at the USC Exposure Analytics lab, they were
handled by trained staff. Pump data were downloaded, checked for errors, and securely stored. Filters were
removed from the PEMs, allowed to equilibrate within a dedicated chamber and gravimetrically weighed

43
in temperature and relative humidity-controlled glove box using an MT-5 calibrated microbalance to obtain
PM 2.5 mass concentration.
Birthweight
Birthweight in grams was abstracted from electronic medical records (EMR) for 210 mothers. Four
mothers did not have birthweight recorded, possibly due to being lost to follow-up, and were removed from
the analysis.
Questionnaire and Other Covariate Data
A priori covariates assessed in this analysis included factors related to maternal demographics,
pregnancy and birth outcomes, meteorology, and study design variables, including recruitment site.
Covariate data was collected during follow-up within the MADRES cohort from a series of in-person and
telephone staff-administered questionnaires in either English or Spanish, ascertained throughout the study
period up until date of infant birth. Anthropometric measurements were conducted via regular clinic visits.
Data from the 3
rd
trimester visit was primarily used to coincide with the exposure period being studied.
Additional covariate data came from the participants’ 1
st
visit, such as race/ethnicity, pre-pregnancy Body
Mass Index (BMI, kg/m
2
), etc. and from pregnancy outcome data, including infant sex.
Maternal demographic variables analyzed for potential confounding included: age at baseline
(years), pre-pregnancy BMI (continuous), education level (completed <12
th
grade, completed high school,
some college, completed college), household income (less than $15,000, $15,000 – 29,999, $30,000 –
49,999, $50,000+, Don’t know), personal smoking status during pregnancy (yes/no), smoking status
(ever/never), diabetes status (no diabetes, glucose-intolerant, gestational diabetes, chronic diabetes),
preeclampsia status (no hypertension, preeclampsia, chronic hypertension, chronic hypertension
w/preeclampsia, gestational hypertension), and total weight gain (kg) during pregnancy. Diabetes and
preeclampsia status were ascertained from EMR, while pre-pregnancy BMI was calculated using self-
reported pre-pregnancy weight and standing height measured by MADRES staff at the first study visit via
stadiometer (Perspectives Enterprises model PE-AIM-101, Portage, MI), or height from EMR if missing

44
from first visit data. Self-reported pre-pregnancy weight was used because initial study visits ranged in
terms of participants’ gestation. Race was recategorized from the NIH categories to a three-level variable
containing Hispanic, Black non-Hispanic, and Other non-Hispanic. This was conducted to save degrees of
freedom in the later regression analysis and because this sample is composed of largely Hispanic women
(81%), followed by a smaller subset of Black non-Hispanic women (11%), and with non-Hispanic Whites,
Asians, and Others combined making up just 8%.
Pregnancy and birth outcome-related potential covariates included: sex of infant (male/female),
parity (defined as 1 or more previous births), and gestational age (GA; weeks). Infant sex was obtained via
EMR, or if missing, through interviewer-administered questionnaires at the 7-14 day post-pregnancy
follow-up. A missing category was created for 6 participants with missing parity. Gestational age at birth
was estimated using a hierarchy of methods including, the preferred ultrasound measurement of crown-
rump length at <14 weeks gestation (60%), ultrasound measurement of fetal biparietal diameter at < 28
weeks gestation (30%), and from physicians’ clinical estimate from EMR (10%).
Meteorological parameters included ambient air temperature (Celsius) (calculated as the average
of minimum and maximum air temperature) and relative humidity (%), both integrated over the 48-hour
sampling period and estimated at the residential location based on a high-resolution (4km x 4km) gridded
surface meteorological dataset (Abatzoglou, 2011). Season was categorized as Cool (Winter), Warm
(Summer), and Transition (Spring and Autumn).
Finally, variables describing home ventilation, time-activity patterns, and the presence of indoor
sources of PM 2.5 came from two different questionnaires. The first was from the 3
rd
trimester visit that asked
questions related to the past month (or since the last visit in the 2
nd
trimester), while the second was from
the exit survey administered after completing the 48-hour personal monitoring period. These variables were
chosen since they correlate with the potential of outdoor PM 2.5 infiltration into the indoor home environment
where participants likely spend most of their time, exposure to outdoor PM 2.5 by spending time outdoors,
or exposure to PM 2.5 generated indoors from sources like cooking or candle use, respectively.

45
To describe these potential relationships in more detail, greater time spent indoors generally
corresponds to greater exposure to indoor PM 2.5, which in turn is predominantly composed of PM 2.5 from
indoor sources (or of indoor origin) and PM 2.5 from outdoor origin (infiltrated indoors). The degree to which
PM 2.5 originating outdoors infiltrates into the indoor home environment depends on several factors
including home ventilation (e.g., AC use, window opening, etc.) (Breen et al., 2014; Habre et al., 2014).
Overall, greater time participants spend outdoors corresponds to potentially greater contribution of outdoor
PM 2.5 to their personal exposures (and vice versa). Additionally, several studies reported air tightness can
be lower (higher leakiness) and air exchange rates can be higher in multi-unit residences (compared to
single homes), which could mean greater potential for PM 2.5 of outdoor origin or from neighboring units
(e.g., secondhand smoke) to infiltrate indoors (King et al., 2010; Price et al., 2006; Rosofsky et al., 2019),
but this likely varies across different contexts. AC use in the home can also remove indoor PM 2.5 or correlate
with lower infiltration of outdoor PM 2.5 (due to more time with windows and doors closed and greater home
sealing to the outdoors).
The final list of variables included: home type (building type/number of attached units), home
ventilation (e.g. AC use, window opening time), time-activity patterns (e.g. time spent indoors, time spent
outdoors), and indoor sources (e.g. cooking smoke, candle and incense smoke). All variables in this final
list were available in both the exit survey and 3
rd
trimester questionnaire, apart from home type, which was
only available from the 3
rd
trimester questionnaire, and candle smoke exposure, which was only asked in
the exit survey. Several of these variables were re-categorized, when necessary, based on the distribution
of the variable (Table S.2.1.).
Statistical Analysis
Descriptive Statistics
Descriptive statistics for birthweight and total personal PM 2.5 were calculated by sample population
characteristics. This preliminary bivariate analysis was also used to elicit potential confounders in our
analysis. The distribution of PM 2.5 exposure and birthweight were assessed to identify any deviations from

46
normality and potential outliers. Differences in birthweight and total personal PM 2.5 by the categorical
sample characteristics were evaluated using analysis of variance (ANOVA) tests. Pearson’s correlation
coefficients were calculated between continuous population characteristics and birthweight and total
personal PM 2.5 separately. Next, a correlation analysis was conducted to assess whether potential covariates
were related to each other to examine collinearity and inform covariate inclusion in our models. Finally, a
chi-square test was conducted to determine how well the two questionnaire measures correlated with one
another for similar variables thereby providing a consistency check for differently worded questions, or
questions that were asked at different points in time and referred to somewhat different time windows (e.g.,
past 48-hour monitoring period versus the last month in the 3
rd
trimester).
Personal vs. Outdoor Residential PM 2.5 Exposure
To assess the relationship between total personal and outdoor PM 2.5 exposure, daily outdoor
residential PM 2.5 concentration was estimated using inverse distance-weighted spatial interpolation from
regulatory monitoring data. Daily estimates were averaged to correspond to the 48-hour monitoring
period and to the 3
rd
trimester of pregnancy. Descriptive statistics were obtained for the same 48-hour
monitoring period and for the 3
rd
trimester, and Pearson’s correlation coefficients were used to evaluate
the relationship between personal and outdoor residential PM 2.5.
Multiple Linear Regression Models
Multiple linear regression models were used to investigate the association between in-utero
exposure to PM 2.5 and the continuous outcome birthweight. All parameter estimates for continuous variables
were reported per 1 SD increase in personal PM 2.5 concentrations, which is equivalent to 17.1 μg/m
3
as
shown in Table 2.1. Maternal age and race/ethnicity were included in all models due to their importance
and inclusion in prior research. Additionally, due to the design of MADRES, recruitment site was also
assessed in our analysis but did not impact findings, so was not included. Our list of potential covariates
based on the previous literature into the effect of air pollution and birth outcomes, and from the bivariate
analysis conducted within this analysis, were assessed for inclusion into our model one-by-one based on

47
evidence of confounding. Confounding was observed by gestational age, parity, diabetes status, infant sex,
and smoking status. Pre-pregnancy BMI and total weight gain during pregnancy also introduced
confounding; however, they were highly correlated with each other and with diabetes status. Each of these
variables were assessed one at a time with the other included covariates and diabetes status was finally
chosen to remain as it impacted the personal PM 2.5 effect estimate the largest of the three.
The final fully adjusted model included the following covariates: GA at birth, maternal age,
race/ethnicity, infant sex, parity, diabetes status, temperature, and personal smoking history. This model
was used to 1) evaluate the effect of total personal PM 2.5 exposure on birthweight, 2) evaluate whether the
effect of total personal PM 2.5 exposure on birthweight was modified by the degree of which indoor vs
outdoor sources contributed to or impacted personal PM 2.5 exposures (broadly derived using questionnaire
variables). The a priori significance level for our adjusted main exposure/outcome analysis was an alpha
of 0.05. Model diagnostics were conducted to ensure they satisfied modeling assumptions and were not
affected by multicollinearity or influential points. Non-linear PM 2.5 effects were evaluated using graphical
plots and by adding polynomials into the model and evaluating statistical significance compared to linear
terms. The analysis was conducted using SAS v9.4 (SAS Institute, Inc., Cary, NC, USA.)
Effect Modification Analyses for PM 2.5 Impacted by Indoor vs Outdoor Sources
As described earlier, the second aim was to evaluate how the effects of total personal PM 2.5
exposure differed when the contribution of outdoor sources (or PM 2.5 of outdoor origin) was higher
compared to indoor sources. Indoor vs outdoor origin of PM 2.5 was approximated using interaction terms
with variables that correspond to time-activity patterns (e.g., time spent indoors vs outdoors), indoor sources
(e.g. cooking, candle use), home ventilation (e.g. AC use, window use), and home type (e.g. building
type/number of attached units). We investigated effect modification by adding an interaction term to the
fully adjusted model, using an a prior significance level of 0.10 for the interaction.

48
Sensitivity Analysis
Several analyses were conducted to evaluate the sensitivity of results to various inclusions. First,
in the fully adjusted model, we examined associations only among full-term births (37 weeks or older
gestation) to assess whether the pre-term births impacted the associations seen in the full sample.
Additionally, due to concerns regarding bias introduced by adjusting for GA, namely, that gestational age
may be a mediator (Wilcox et al., 2011), we performed the analysis without adjustment for GA. Finally, a
model was run without the inclusion of the highest 4 personal PM 2.5 concentrations, determined by
concentrations being 2 SDs greater than the mean, to elicit their leverage on results.
RESULTS
Descriptive Statistics
Of the 214 mothers who participated in the personal exposure monitoring study, nine participants
were removed due to incomplete or erroneous personal PM 2.5 exposure data or birth outcomes data,
resulting in a final analytical sample of 205 mother-infant dyads (Table 2.1.). The women in the study were
predominantly Hispanic (81%) and lower income, with over 55% of women reporting a household income
less than $50,000 a year. Additionally, around 67% of participants were overweight or obese prior to
pregnancy and most had at least one prior pregnancy (63%). Birthweight was normally distributed with a
mean (SD) of 3,291.2 (485.1) grams. Total personal PM 2.5 exposure was right skewed with a mean (SD) of
22.3 (17.1) μg/m
3
and median (IQR) of 18.2 (14.3) μg/m
3
. The participants had a mean (SD) age of 28.2
(6.0) years, delivered at a mean gestational age at birth of 39.1 (1.5) weeks, and gained on average 10.9
(6.9) kilograms throughout pregnancy.
Table 2.1. Descriptive Statistics of Study Participants (N=205).
Variable Mean (SD)
or n (%)
Variable Mean (SD)
or n (%)
Personal PM 2.5 (μg/m
3
) 22.3 (17.1) Pre-Pregnancy BMI (kg/m
2
) 28.8 (6.8)
Birthweight (g) 3,291.2 (485.1) Normal 62 (30.2%)
Total Weight Gain (kg) 10.9 (6.9) Overweight 64 (31.2%)
Gestational Age (weeks) 39.1 (1.5) Obese 79 (38.5%)
Maternal Age (years) 28.2 (6.0) Parity

49
Gender Yes 130 (63.4%)
Female 105 (51.2%) No 69 (33.7%)
Male 100 (48.8%) Missing 6 (2.9%)
Race Maternal Income
Hispanic 166 (81.0%) $50,000-$99,999 14 (6.8%)
Black, Non-Hispanic 23 (11.2%) $30,000-$49,999 29 (14.1%)
Other, Non-Hispanic 16 (7.8%) $15,000-$29,999 46 (22.4%)
Education Less than $15,000 41 (20.0%)
<12th grade 49 (23.9%) Don't know 75 (36.6%)
Completed High School 64 (31.2%) Smoking
Some College 62 (30.2%) Ever 43 (21.0%)
Completed College 30 (14.6%) Never 162 (79.0%)
Diabetes Temperature (°C) 19.3 (4.2)
Normal 136 (66.3%) Season

Glucose Intolerant 46 (22.4%) Warm 40 (19.5%)
Gestational Diabetes 10 (4.9%) Cool 64 (31.2%)
Chronic Diabetes 13 (6.3%) Transition 101 (49.3%)
Notes: PM2.5 = particulate matter with aerodynamic diameter less than 2.5µm; BMI = body mass index; SD = standard
deviation; g = grams; kg = kilograms; transition = spring and autumn.

Sociodemographic and Household Characteristics in Relation to Birthweight
Mothers who spent most or all of the time indoors had infants with significantly higher birthweight
(most and all of the time: 3,332.0g vs. none and a little of the time: 3,065.3g; p = 0.005), while those who
answered yes to using AC during the sampling period had infants that were about 190 grams greater in
birthweight than mothers that did not use AC (p = 0.013). Participants who had at least one child prior to
this pregnancy had infants with higher birthweight (yes: 3,338.8g vs. no: 3,180.6g; p = 0.042). Infants of
women who have completed at least college, had gestational or chronic diabetes, or were in the non-
Hispanic Other category, had higher birthweight compared to their counterparts; however, none of these
differences met statistical significance (Table S.2.1.).
Birthweight displayed a positive correlation with gestational age (Pearson r = 0.43; p <0.001) and
total weight gain throughout pregnancy (r = 0.29; p <0.001), while maternal age showed no correlation (r
= 0.02; p = 0.738; Table S.2.2.).

50
Sociodemographic and Household Characteristics in Relation to Total Personal PM 2.5 Exposure
A statistically significant difference in personal PM 2.5 was observed by maternal income; however,
no obvious pattern emerged, with the highest and lowest income groups having the highest personal PM 2.5
exposure (p = 0.025; Table S.2.1.). Participants who opened their windows none or a little of the time
during the sampling period had slightly higher personal PM 2.5 exposure (24.7μg/m
3
) vs. most and all of the
time (20.2μg/m
3
), which was marginally significant (p = 0.058). Personal PM 2.5 differed by season of
sampling (warm: 19.3μg/m
3
vs. transition: 21.0μg/m
3
vs. cool: 26.3; p = 0.072). Additionally, personal
PM 2.5 was significantly negatively associated with average 3
rd
trimester temperature (r = -0.15; p = 0.038;
Table S.2.2.). Next, women who spent most or all of the time inside during the sampling period had lower
personal PM 2.5 exposure (21.4 vs 25.3 μg/m
3
for women who spent none or a little of the time indoors; p =
0.249). Using the monitoring time-aligned exit survey question on cooking smoke exposure, there was no
significant difference between those reporting being near cooking smoke and those that did not (none of
the time: 22.1μg/m
3
vs. a little, most, or all of the time 21.8μg/m
3
; p = 0.884).
Relationship Between Personal and Outdoor PM 2.5 Exposure
The mean (SD) personal PM 2.5 was 22.3 (17.1) μg/m
3
, while the outdoor residential estimate had a
mean of 11.9 (5.5) μg/m
3
for the same 48-hour monitoring period, and 12.0 (2.3) μg/m
3
for the 3
rd
trimester.
Figure 2.1. depicts these relationships between total personal and outdoor residential PM 2.5. During the
monitoring period, there was statistically significant yet weak correlation between total personal PM 2.5 and
ambient PM 2.5 (r = 0.19; p = 0.006). A weak, positive and non-significant correlation between total personal
PM 2.5 and 3
rd
trimester ambient PM 2.5 was also observed (r = 0.11; p = 0.110).

51
Figure 2.1. Relationship of Personal PM 2.5 and Outdoor PM 2.5 in (a) the 48-hour Monitoring Period and
(b) the Third Trimester of Pregnancy.
(a) (b)

Association of Total Personal PM 2.5 Exposure with Birthweight
We found no significant association between PM 2.5 and birthweight (β=37.4; 95% CI: -29.6, 104.3;
p = 0.273, per 1 SD increase in PM 2.5) in the crude (unadjusted) regression model. Results remained similar
in the fully-adjusted model (with maternal age, GA, maternal race/ethnicity, infant sex, parity, diabetes
status, smoking status, and 3
rd
trimester average temperature, (β=38.6; 95% CI: -21.1, 98.2; p = 0.204), as
shown in Table 2.2. In the fully adjusted model, a one week increase in GA was associated with a 180.3g
increase in birthweight (p <0.001), females were on average 124.8g lighter than males (p = 0.033), and
participants that had not had a pregnancy before had on average 323.4g lighter babies compared to those
that had (p <0.001). Finally, diabetes status was also an important predictor of birthweight, with participants
with chronic diabetes (379.4g; p = 0.003) and gestational diabetes (300.9g; p = 0.028) having higher
birthweight infants compared to those without diabetes.
Table 2.2. Regression Results for Base Model of PM 2.5 and Birthweight (N = 205).
Variable β 95% CI
Lower
95% CI
Upper
p-value
Intercept -
3452.3
-5096.7 -1807.9 <0.001
Personal PM 2.5 (μg/m
3
)
a
38.6 -21.1 98.2 0.204
Gestational Age (weeks) 180.3 141.1 219.5 <0.001
Maternal Age (years) -8.8 -19.8 2.3 0.121
Temperature (°C) ^ 11.3 -5.8 28.4 0.194

52
Race/ethnicity

Hispanic -135.3 -356.9 86.3 0.230
Black, non-Hispanic -248.7 -520.0 22.6 0.072
Other, non-Hispanic Ref.
Sex of Infant

Female -124.8 -239.2 -10.3 0.033
Male Ref.
Parity

Missing 176.4 -166.0 518.7 0.311
No -323.4 -459.4 -187.4 <0.001
Yes Ref.
Diabetes

Chronic Diabetes 379.4 129.7 629.1 0.003
Gestational Diabetes Mellitus
300.9 33.1 568.7 0.028
Glucose Intolerant 111.3 -26.9 249.5 0.114
Normal Ref.
Smoking

Ever Smoker -175.8 -319.7 -31.8 0.017
Never Smoker Ref.
Notes:
a
Per 1 SD increase in personal PM2.5; PM2.5 = particulate matter with aerodynamic diameter less
than 2.5µm; CI = confidence interval; ^ - third trimester average temperature in degrees Celsius; Ref. =
reference level.
Effect Modification of Total Personal PM 2.5 by Contribution of Indoor vs Outdoor Sources
While total personal PM 2.5 was not associated with birthweight in our first aim, this association
differed significantly by several factors (Table 2.3.). Home type was a significant effect modifier of
personal PM 2.5 exposure on birthweight (interaction p = 0.028, Figure 2.2.b). Participants living in a “house
with no joining walls” (β=156.9; 95% CI: 26.9, 287.0) had a positive association with birthweight; while a
negative association was observed as the number of units in the housing building increased (2-4 units: β=-
16.6; 95% CI:-111.9, 78.7; 5+ units: β=-62.6; 95% CI:-184.9, 59.6). Additionally, the effect of PM 2.5 on
birthweight was significantly different by AC use (interaction p = 0.008), with more negative associations
for participants that reported no AC use on the exit survey (β=-27.6; 95% CI:-101.5, 46.3), compared to
any AC use during the 48-hour monitoring period (β=139.9; 95% CI: 42.9, 237.0) (Figure 2.2.d). A similar
significant interaction and pattern was observed for AC use reported in the 3
rd
trimester (Table 2.3.).
Participants who reported any exposure to smoke from candles or incense (only assessed in exit
survey) had a negative association between PM 2.5 and birthweight (β=-144.7; 95% CI: -282.7, -6.8), vs

53
those who reported no exposure (β=81.2; 95% CI: 15.9, 146.6). This interaction was statistically significant
(p = 0.004). There was no significant interaction with cooking smoke exposure in the 48-hour monitoring
period (none: β=75.5; 95% CI: -3.5, 154.5 vs any: β=-10.6; 95% CI: -100.0, 78.8; interaction p = 0.153).
Results were similar when using the 3
rd
trimester questionnaire (Table 2.3.).
There were also consistent, observable differences in the effect of PM 2.5 on birthweight for variables
related to time-activity patterns, although these interactions were not statistically significant (Table 2.3.).
Figure 2.2.c depicts when participants spent most and all of their time indoors during the 48-hour
monitoring period (at their residence or someone else’s), PM 2.5 was positively associated with birthweight
(β=57.1; 95% CI: -7.3, 121.6), as compared to participants who spent none and a little of their time indoors
(β=-45.1; 95% CI: -208.3, 118.1). The 3
rd
trimester questionnaire revealed a similar pattern, with a positive
effect of PM 2.5 on birthweight when participants spent greater than 16 hours inside per day (β=50.0; 95%
CI: -13.9, 114.0) compared to a slight negative association for participants who spent less than or equal to
16 hours inside (β=-25.6; 95% CI: -184.8, 133.6).
Table 2.3. Estimated Change in Birthweight (g) per 1 SD Increase in Personal PM 2.5 from Interaction
Analyses (N = 204).
Time Activity Pattern β 95% CI
Lower
95% CI
Upper
p-value
Time Spent Indoors
How much of the time did you spend indoors (at your residence, or someone else's
residence)?
a

0.255
None and a little of the time -45.1 -208.3 118.1
Most and all of the time 57.1 -7.3 121.6
Thinking back to a typical weekday in this past week, approximately how
many hours (out of 24 hours in total) did you spend indoors?
b

0.383
≤16 hours -25.6 -184.8 133.6
>16 hours 50.0 -13.9 113.9
Time Spent Outdoors
How much of the time did you spend outdoors (not commuting in a car, bus
or train)?
a

0.402
None and a little of the time 59.5 -13.6 132.6
Most and all of the time 6.8 -94.2 107.7
Thinking back to a typical weekday in this past week, approximately how
many hours (out of 24 hours in total) did you spend outdoors?
b

0.411

54
< 8 hours 46.0 -16.5 108.5
≥ 8 hours -33.5 -215.2 148.3
Home Characteristics and Ventilation
Home Type
Which best describes the home in which you currently live most of the
time?
b

0.028
A single-family house (no joining wall) 156.9 26.9 287.0
A building with 2-4 attached Units -16.6 -111.9 78.7
A building with 5+ attached Units -62.6 -184.9 59.6
Missing 145.4 -4.1 294.9
Time With Windows Open
How much of the time were windows (or porch/balcony doors if applicable)
open in your home, when you were there with the sampler?
a

0.936
None and a little of the time 33.8 -35.3 102.9
Most and all of the time 28.2 -89.3 145.8
On average, how much of the time were the windows open in your home this
past week?
b

0.230
None and a little of the time 53.9 -14.2 122.0
Most and all of the time -30.2 -151.8 91.5
Air Conditioner Use
How much of the time was the air conditioner used in your home, when you
were there with the sampler?
a

0.008
None of the time -27.6 -101.5 46.3
A little, most, and all of the time 139.9 42.9 237.0
Do you use air conditioning in your home?
b
0.044
No -24.3 -107.1 58.5
Yes 99.4 13.5 185.3
Indoor Sources
Cooking
How much of the time were you close to smoke or fumes from cooking
(yourself, or nearby cooking by someone else) e.g. burnt toast, barbeque, stir
fry, etc.?
a

0.153
None of the time 75.5 -3.5 154.5
A little, most, and all of the time -10.6 -100.0 78.8
Since we last saw you/spoke to you in your first/second trimester, on
average, how many times a week do you cook (using the stove/range/oven,
not microwave)?
b

0.085
Never 158.0 9.2 306.8
1 or more times a week 15.2 -50.2 80.5
Candle or Incense
How much of the time were you close to smoke from candles or incense
burning nearby?
a

0.004
None of the time 81.2 15.9 146.6
A little, most, and all of the time -144.7 -282.7 -6.8

55
Notes: All interactions are adjusted for gestational age, gender, parity, race/ethnicity, maternal age, diabetes status, smoking, and
temperature; PM2.5 = particulate matter with aerodynamic diameter less than 2.5µm; CI = confidence interval;
a
From exit
questionnaire administered to participants after completing the 48-hour personal exposure monitoring period;
b
From 3
rd
trimester
survey; bolded = statistically significant at p-value < 0.1.

56
Figure 2.2. Predicted Relationship of Personal PM 2.5 Exposure on Birthweight (a) Overall, (b) by Type
of Home, (c) Time Spent Indoors, and (d) Air Conditioner Use at Home, for Each Level of the
Interaction Variable Where Applicable.
(a) Overall (b) By Home Type

(c) By Time Spent Indoors (d) By Air Conditioner Use

Sensitivity Analyses Results
When the fully adjusted model was restricted to just those participants that had a full-term birth
(≥37 weeks gestation; n=182), the effect of total personal PM 2.5 on birthweight increased slightly (β=55.0;
95% CI: -6.2, 116.1), compared to the base model used in aim 1 (β=38.6; 95% CI: -21.1, 98.2; P = 0.204).
In non-diabetics only (n=181), no association between total personal PM 2.5 on birthweight was observed
(β=19.2; 95% CI: -44.7, 83.2). When GA was not adjusted for, there was a 24% attenuation in the effect
estimate for total personal PM 2.5 on birthweight (β=29.3; 95% CI: -41.8, 100.5). Finally, after excluding
three observations that had high leverage and particularly high PM 2.5 (>95μg/m
3
) from the model, the effect
of total personal PM 2.5 on birthweight changed direction but remained non-significant (β=-40.1; 95% CI: -
122.3, 42.1).

57
DISCUSSION
Using data from the MADRES in-utero personal exposure monitoring study, we evaluated the
effect of total personal PM 2.5 in the 3
rd
trimester on birthweight in a largely lower income, Hispanic
population in Los Angeles, CA. To our knowledge, this is the first time this has been attempted in a health
disparities population. Our study finds that total personal PM 2.5 was not statistically significantly associated
with birthweight. Most studies of outdoor air pollution generally found a slight negative association with
birthweight (Stieb et al., 2012); however, those studies were aimed at investigating personal exposure to
PM 2.5 of outdoor origin rather than total personal PM 2.5 and generally relied on outdoor estimates of PM 2.5
as its surrogate. Albeit a different question to the effect of total personal PM 2.5 on birthweight, these outdoor
estimates generally fail to account for time-activity patterns and infiltration of outdoor pollution into the
home, thereby likely suffer from measurement error.
One study did look at the effect of total personal PM 2.5 on birthweight in the 2
nd
trimester using
personal monitoring over a 48-hour period in a cohort of non-smoking women in Poland. They found an
increase of ~30 μg/m³ in PM 2.5 was associated with a 97.2g (95% CI: - 201.0, 6.6) decrease in birth weight
(Jedrychowski et al., 2009). Despite the Poland findings not being statistically significant, a possible
hypothesis for the difference in findings could be related to the differences between the women participating
in the two studies. For example, compared to our study, participants were free from chronic diseases
including diabetes, which we found to be a significant predictor of higher birthweight in our study.
Additionally, sources and chemical components of PM 2.5 exposure in Poland may be different compared to
our study area in urban Los Angeles, CA. Studies have shown that PM 2.5 sources and chemical components
can vary in their effect on birthweight, with differences observed across regions (Basu et al., 2014; Bell et
al., 2007), and across race/ethnic groups, especially in California (Bell and Ebisu, 2012). Our study
participants were largely Hispanic from Los Angeles County, while the Polish study population was
predominantly non-Hispanic Whites (Jedrychowski et al., 2009). This highlights the importance of treating
PM 2.5 as a mixture in health analyses, with variable contributions from a wide range of indoor and outdoor
sources with potentially differing physiochemical properties, components, and effects on birthweight.

58
Comparing outdoor PM 2.5 effects across regions, or outdoor to total personal PM 2.5 effects, does not
necessarily take into account this complexity or heterogeneity.
While there was no association between total personal PM 2.5 and birthweight, we did find evidence
that home characteristics, such as home type and AC use, as well as exposure to candle or incense smoke,
modified this association. Mothers residing in multi-unit buildings had a negative association of personal
PM 2.5 with birthweight, compared to a strong positive association for those who reside in a single-family
home. One possible reason for this is that individuals in multi-unit homes may have greater second-hand
smoke infiltration into their home from neighboring units (King et al., 2010; Price et al., 2006), and second-
hand smoke has been shown to be negatively associated with birthweight (Ghosh et al., 2013; Wahabi et
al., 2013). We also considered whether single-family home type could be acting as a proxy for higher
income. However, maternal income and educational attainment did not correlate with home type in our
sample (results not shown). Additionally, despite having similar total personal PM 2.5 exposure, the effect of
PM 2.5 on birthweight was significantly lower for participants that did not use AC compared to those that
did. Using AC at home likely correlates with greater sealing of the home or closing of windows and doors
to operate the AC unit(s), which also correlates with less infiltration outdoor PM 2.5 indoors (and thus less
exposure to PM 2.5 of outdoor origin).
Although we did not have complete information on all the possible indoor sources of PM 2.5, we
saw evidence of significantly more negative or potentially harmful effects of candle and/or incense burning
indoors on birthweight. Previous literature has shown that prenatal incense burning was associated with
lower birthweight (Chen and Ho, 2016). One possible explanation is candle and incense burning emit black
carbon (BC) and other chemicals indoors (Habre et al., 2014; Pagels et al., 2009; Stabile et al., 2012).
Several studies have reported an association between BC concentrations in PM 2.5 and low birthweight;
however, they were using outdoor BC as a surrogate or marker of outdoor, traffic-related air pollution
(Lakshmanan et al., 2015; Slama et al., 2007). Bove et al. (2019) found BC accumulated on the fetal side
of the human placenta, representing a potential mechanism for negative health effects. However, without
further information on chemical composition and properties of the personal PM 2.5 exposure mixture in these

59
studies, it is difficult to conclude whether the candles and incense burning mixture as a whole or any
particular component of it, such as BC, is driving these adverse effects.
We did not find consistent effect modification results for exposure to PM 2.5 from cooking, as
another potentially important indoor source in our population. Most cooking smoke exposure and
birthweight studies have concentrated on solid fuel sources (e.g., coal, wood), often in low- and middle-
income countries, but generally find a negative association with birthweight (Wylie et al., 2017; Zhang et
al., 2016). These findings may not be a suitable comparison with our study sample considering our
participants were based in Los Angeles, CA where solid fuel cooking is not common, and where the
composition or mixture of cooking related exposures may be different due to most participants using gas
stoves (data not shown). It is also possible that cooking emits particles in the ultrafine size range (<0.1µm
in aerodynamic diameter) which do not contribute significantly to PM 2.5 mass concentrations, and thus, our
measurements may not be sensitive enough to differentiate their contribution (as compared to particle
number concentrations for example, which were not available in our study).
Our analyses also revealed a consistent pattern where personal PM 2.5 exposure with greater
influence or contribution of outdoor sources was generally more strongly associated with lower birthweight,
despite these interactions not reaching statistical significance. This was true for greater time spent outside
(and less time spent inside), and greater time with open windows. However, these associations should be
explored further, as the effect modification of time with windows open reported on the exit survey was less
pronounced. While our results may be underpowered to tease apart these differences, our results are
consistent with prior studies assessing the impact of specific outdoor sources of PM 2.5 or their surrogates
such as on road gasoline and on road diesel, or residing closer to major roadways, respectively, which were
associated with greater risk of LBW compared to PM 2.5 as a whole (Bell et al., 2010; Laurent et al., 2016).
There are several strengths of this analysis. The first is the use of personal monitoring that provides
a unique opportunity to examine personal exposure to PM 2.5 and disentangle its impact on birthweight when
it was more impacted by outdoor sources. This approach drastically reduces exposure measurement error
as compared to using outdoor estimates of PM 2.5 despite it being limited to a small sample (Gray et al.,

60
2010). Next, we were able to evaluate total personal PM 2.5 exposure in the 3
rd
trimester, which may be
particularly important for birthweight, given that most fetal growth occurs late in pregnancy. Most studies
on the effect of PM 2.5 on birthweight have used ambient monitoring data to estimate personal exposure to
PM 2.5 of outdoor origin, rather than the total personal PM 2.5 to which individuals are exposed.
Understanding the effects of outdoor PM 2.5 on health is certainly an important question to evaluate, due to
this being the fraction of PM 2.5 that is regulated, but it is not the same question as the effect of total personal
PM 2.5 which takes into account all indoor, outdoor, and personal activity related sources that contribute to
personal exposure as a result of realistic day-to-day behaviors and time-activity patterns. However, we were
also able to indirectly investigate whether the effects of personal PM 2.5 differed when it was more impacted
by indoor vs outdoor sources using interaction analyses with detailed, time-aligned questionnaire variables.
It is also important to note that the chemical composition and size distribution of outdoor PM 2.5 changes as
it infiltrates indoors, which further highlights the importance of deciphering the independent effects of
personal exposure to PM 2.5 of indoor versus outdoor origin (Meng et al., 2007).
Furthermore, the MADRES cohort study is a well characterized prospective study in a health
disparities population, with a host of covariates available, making this an ideal study to assess the research
question at hand. We also had the advantage of using two questionnaire data sources that differed in their
time coverage and alignment (a longer-term 3
rd
trimester questionnaire vs an exit survey immediately
following the 48-hour monitoring period). This also allowed us to shed light on whether the 48-hour
sampling period reasonably represented behaviors, time-activity patterns, and 3
rd
trimester exposures in
general.
The sample size of this study is a potential limitation with a final working sample of 205
participants, which while small for population-based studies is actually reasonably large for personal
exposure monitoring studies (Dadvand et al., 2012; Sarnat et al., 2000; Suh and Zanobetti, 2010). Despite
this limitation, we were still able to observe differences in the influence of personal PM 2.5 on birthweight
by factors that drive indoor/outdoor source contributions, and most significantly for AC use and home type.
Finally, participation bias may be a factor regarding who from the MADRES cohort chose to participate in

61
the personal exposure monitoring study; however, participants who chose to participate were not noticeably
different than the larger MADRES cohort study apart from being slightly more likely to have had a prior
child (data not presented).
CONCLUSION
Overall, our results did not find a significant association between total personal PM 2.5 exposure and
birthweight; however, we did find evidence that multi-unit housing (vs. single family homes), candle and/or
incense smoke exposure, and greater outdoor source contributions to personal PM 2.5 were more strongly
associated with lower birthweight. This highlights the importance of disentangling the mixture and
apportioning PM 2.5 by sources for health analyses, including potentially a more refined or chemically
speciated approach to apportion indoor from outdoor source contributions to personal PM 2.5.

62
REFERENCES
Abatzoglou, J.T., 2011. Development of gridded surface meteorological data for ecological applications
and modelling - Abatzoglou - 2013 - International Journal of Climatology - Wiley Online
Library. International Journal of Climatology 33, 121–131.
Barnes, S.K., Ozanne, S.E., 2011. Pathways linking the early environment to long-term health and
lifespan. Progress in Biophysics and Molecular Biology, Developmental programming of
phenotype diversity: Epigenetics and disease risk 106, 323–336.
https://doi.org/10.1016/j.pbiomolbio.2010.12.005
Bastain, T.M., Chavez, T., Habre, R., Girguis, M.S., Grubbs, B., Toledo-Corral, C., Amadeus, M.,
Farzan, S.F., Al-Marayati, L., Lerner, D., Noya, D., Quimby, A., Twogood, S., Wilson, M.,
Chatzi, L., Cousineau, M., Berhane, K., Eckel, S.P., Lurmann, F., Johnston, J., Dunton, G.F.,
Gilliland, F., Breton, C., 2019. Study Design, Protocol and Profile of the Maternal And
Developmental Risks from Environmental and Social Stressors (MADRES) Pregnancy Cohort: a
Prospective Cohort Study in Predominantly Low-Income Hispanic Women in Urban Los
Angeles. BMC Pregnancy Childbirth 19. https://doi.org/10.1186/s12884-019-2330-7
Basu, R., Harris, M., Sie, L., Malig, B., Broadwin, R., Green, R., 2014. Effects of fine particulate matter
and its constituents on low birth weight among full-term infants in California. Environmental
Research 128, 42–51. https://doi.org/10.1016/j.envres.2013.10.008
Bell, M.L., Belanger, K., Ebisu, K., Gent, J.F., Lee, H.J., Koutrakis, P., Leaderer, B.P., 2010. Prenatal
Exposure to Fine Particulate Matter and Birth Weight. Epidemiology 21, 884–891.
https://doi.org/10.1097/EDE.0b013e3181f2f405
Bell, M.L., Dominici, F., Ebisu, K., Zeger, S.L., Samet, J.M., 2007. Spatial and Temporal Variation in
PM2.5 Chemical Composition in the United States for Health Effects Studies. Environ Health
Perspect 115, 989–995. https://doi.org/10.1289/ehp.9621
Bell, M.L., Ebisu, K., 2012. Environmental Inequality in Exposures to Airborne Particulate Matter
Components in the United States. Environ Health Perspect 120, 1699–1704.
https://doi.org/10.1289/ehp.1205201
Breen, M.S., Schultz, B.D., Sohn, M.D., Long, T., Langstaff, J., Williams, R., Isaacs, K., Meng, Q.Y.,
Stallings, C., Smith, L., 2014. A review of air exchange rate models for air pollution exposure
assessments. J Expo Sci Environ Epidemiol 24, 555–563. https://doi.org/10.1038/jes.2013.30
Carroll, R.J., 2005. Measurement Error in Epidemiologic Studies, in: Encyclopedia of Biostatistics.
American Cancer Society. https://doi.org/10.1002/0470011815.b2a03082
Chen, L.-Y., Ho, C., 2016. Incense Burning during Pregnancy and Birth Weight and Head Circumference
among Term Births: The Taiwan Birth Cohort Study. Environ Health Perspect 124, 1487–1492.
https://doi.org/10.1289/ehp.1509922
Choi, H., Rauh, V., Garfinkel, R., Tu, Y., Perera, F.P., 2008. Prenatal exposure to airborne polycyclic
aromatic hydrocarbons and risk of intrauterine growth restriction. Environmental Health
Perspectives 116, 658–665. https://doi.org/10.1289/ehp.10958
Choi, H., Wang, L., Lin, X., Spengler, J.D., Perera, F.P., 2012. Fetal Window of Vulnerability to
Airborne Polycyclic Aromatic Hydrocarbons on Proportional Intrauterine Growth Restriction.
PLOS ONE 7, e35464. https://doi.org/10.1371/journal.pone.0035464
Clemente, D.B.P., Casas, M., Vilahur, N., Begiristain, H., Bustamante, M., Carsin, A.-E., Fernández,
M.F., Fierens, F., Gyselaers, W., Iñiguez, C., Janssen, B.G., Lefebvre, W., Llop, S., Olea, N.,
Pedersen, M., Pieters, N., Santa Marina, L., Souto, A., Tardón, A., Vanpoucke, C., Vrijheid, M.,
Sunyer, J., Nawrot, T.S., 2016. Prenatal Ambient Air Pollution, Placental Mitochondrial DNA
Content, and Birth Weight in the INMA (Spain) and ENVIRONAGE (Belgium) Birth Cohorts.
Environ Health Perspect 124, 659–665. https://doi.org/10.1289/ehp.1408981
Dadvand, P., de, N.A., Triguero, -Mas Margarita, Schembari, A., Cirach, M., Amoly, E., Figueras, F.,
Basaga, ña X., Ostro, B., Nieuwenhuijsen, M., 2012. Surrounding Greenness and Exposure to

63
Air Pollution During Pregnancy: An Analysis of Personal Monitoring Data. Environmental
Health Perspectives 120, 1286–1290. https://doi.org/10.1289/ehp.1104609
Dadvand, P., Ostro, B., Figueras, F., Foraster, M., Basagaña, X., Valentín, A., Martinez, D., Beelen, R.,
Cirach, M., Hoek, G., Jerrett, M., Brunekreef, B., Nieuwenhuijsen, M.J., 2014. Residential
Proximity to Major Roads and Term Low Birth Weight: The Roles of Air Pollution, Heat, Noise,
and Road-Adjacent Trees. Epidemiology 25, 518–525.
https://doi.org/10.1097/EDE.0000000000000107
Dadvand, P., Parker, J., Bell, M.L., Bonzini, M., Brauer, M., Darrow, L.A., Gehring, U., Glinianaia, S.V.,
Gouveia, N., Ha, E., Leem, J.H., van den Hooven, E.H., Jalaludin, B., Jesdale, B.M., Lepeule, J.,
Morello-Frosch, R., Morgan, G.G., Pesatori, A.C., Pierik, F.H., Pless-Mulloli, T., Rich, D.Q.,
Sathyanarayana, S., Seo, J., Slama, R., Strickland, M., Tamburic, L., Wartenberg, D.,
Nieuwenhuijsen, M.J., Woodruff, T.J., 2013. Maternal Exposure to Particulate Air Pollution and
Term Birth Weight: A Multi-Country Evaluation of Effect and Heterogeneity. Environ Health
Perspect 121, 267–373. https://doi.org/10.1289/ehp.1205575
Ebisu, K., Belanger, K., Bell, M.L., 2014. The Association between Airborne PM2.5 Chemical
Constituents and Birth Weight–Implication of Buffer Exposure Assignment. Environ Res Lett 9.
https://doi.org/10.1088/1748-9326/9/8/084007
Ghosh, J.K.C., Wilhelm, M., Ritz, B., 2013. Effects of Residential Indoor Air Quality and Household
Ventilation on Preterm Birth and Term Low Birth Weight in Los Angeles County, California. Am
J Public Health 103, 686–694. https://doi.org/10.2105/AJPH.2012.300987
Gray, S.C., Edwards, S.E., Miranda, M.L., 2010. Assessing Exposure Metrics for PM and Birthweight
Models. J Expo Sci Environ Epidemiol 20, 469–477. https://doi.org/10.1038/jes.2009.52
Gray, S.C., Gelfand, A.E., Miranda, M.L., 2011. Hierarchical spatial modeling of uncertainty in air
pollution and birth weight study. Statistics in Medicine 30, 2187–2198.
https://doi.org/10.1002/sim.4234
Habre, R., Coull, B., Moshier, E., Godbold, J., Grunin, A., Nath, A., Castro, W., Schachter, N., Rohr, A.,
Kattan, M., Spengler, J., Koutrakis, P., 2014. Sources of indoor air pollution in New York City
residences of asthmatic children. J Expo Sci Environ Epidemiol 24, 269–278.
https://doi.org/10.1038/jes.2013.74
Hales, C.N., 1997. Non-insulin-dependent diabetes mellitus. British Medical Bulletin 53, 109–122.
https://doi.org/10.1093/oxfordjournals.bmb.a011594
Harris, G., Thompson, W.D., Fitzgerald, E., Wartenberg, D., 2014. The association of PM2.5 with full
term low birth weight at different spatial scales. Environ Res 134, 427–434.
https://doi.org/10.1016/j.envres.2014.05.034
Huang, C., Nichols, C., Liu, Y., Zhang, Y., Liu, X., Gao, S., Li, Z., Ren, A., 2015. Ambient air pollution
and adverse birth outcomes: a natural experiment study. Popul Health Metr 13.
https://doi.org/10.1186/s12963-015-0050-4
Huxley, R., Owen, C.G., Whincup, P.H., Cook, D.G., Rich-Edwards, J., Smith, G.D., Collins, R., 2007. Is
birth weight a risk factor for ischemic heart disease in later life? The American Journal of Clinical
Nutrition 85, 1244–1250. https://doi.org/10.1093/ajcn/85.5.1244
Jedrychowski, W., Perera, F., Mrozek-Budzyn, D., Mroz, E., Flak, E., Spengler, J.D., Edwards, S., Jacek,
R., Kaim, I., Skolicki, Z., 2009. Gender differences in fetal growth of newborns exposed
prenatally to airborne fine particulate matter. Environmental Research 109, 447–456.
https://doi.org/10.1016/j.envres.2009.01.009
Jornayvaz, F.R., Vollenweider, P., Bochud, M., Mooser, V., Waeber, G., Marques-Vidal, P., 2016. Low
birth weight leads to obesity, diabetes and increased leptin levels in adults: the CoLaus study.
Cardiovasc Diabetol 15. https://doi.org/10.1186/s12933-016-0389-2
King, B.A., Travers, M.J., Cummings, K.M., Mahoney, M.C., Hyland, A.J., 2010. Secondhand Smoke
Transfer in Multiunit Housing. Nicotine Tob Res 12, 1133–1141.
https://doi.org/10.1093/ntr/ntq162

64
Kioumourtzoglou, M.-A., Spiegelman, D., Szpiro, A.A., Sheppard, L., Kaufman, J.D., Yanosky, J.D.,
Williams, R., Laden, F., Hong, B., Suh, H., 2014. Exposure measurement error in PM2.5 health
effects studies: A pooled analysis of eight personal exposure validation studies. Environ Health
13, 2. https://doi.org/10.1186/1476-069X-13-2
Kiserud, T., Benachi, A., Hecher, K., Perez, R.G., Carvalho, J., Piaggio, G., Platt, L.D., 2018. The World
Health Organization fetal growth charts: concept, findings, interpretation, and application.
American Journal of Obstetrics & Gynecology 218, S619–S629.
https://doi.org/10.1016/j.ajog.2017.12.010
Lakshmanan, A., Chiu, Y.-H.M., Coull, B.A., Just, A.C., Maxwell, S.L., Schwartz, J., Gryparis, A.,
Kloog, I., Wright, R.J., Wright, R.O., 2015. Associations between Prenatal traffic-related air
pollution exposure and birth weight: Modification by sex and maternal pre-pregnancy body mass
index. Environ Res 137, 268–277. https://doi.org/10.1016/j.envres.2014.10.035
Lamichhane, D.K., Leem, J.-H., Lee, J.-Y., Kim, H.-C., 2015. A meta-analysis of exposure to particulate
matter and adverse birth outcomes. Environ Health Toxicol 30.
https://doi.org/10.5620/eht.e2015011
Laurent, O., Hu, J., Li, L., Kleeman, M.J., Bartell, S.M., Cockburn, M., Escobedo, L., Wu, J., 2016. Low
birth weight and air pollution in California: Which sources and components drive the risk?
Environ Int 92–93, 471–477. https://doi.org/10.1016/j.envint.2016.04.034
Li, X., Huang, S., Jiao, A., Yang, X., Yun, J., Wang, Y., Xue, X., Chu, Y., Liu, F., Liu, Y., Ren, M.,
Chen, X., Li, N., Lu, Y., Mao, Z., Tian, L., Xiang, H., 2017. Association between ambient fine
particulate matter and preterm birth or term low birth weight: An updated systematic review and
meta-analysis. Environmental Pollution 227, 596–605.
https://doi.org/10.1016/j.envpol.2017.03.055
Li, Zhou, Tang, Y., Song, X., Lazar, L., Li, Zhen, Zhao, J., 2019. Impact of ambient PM2.5 on adverse
birth outcome and potential molecular mechanism. Ecotoxicology and Environmental Safety 169,
248–254. https://doi.org/10.1016/j.ecoenv.2018.10.109
Martin, J.A., Hamilton, B.E., Osterman, M.J.K., Driscoll, A.K., 2018. National Vital Statistics Reports
Volume 68, Number 13, November 30, 2019, Births: Final Data for 2018 47.
Meng, Q.Y., Turpin, B.J., Lee, J.H., Polidori, A., Weisel, C.P., Morandi, M., Colome, S., Zhang, J.,
Stock, T., Winer, A., 2007. How Does Infiltration Behavior Modify the Composition of Ambient
PM2.5 in Indoor Spaces? An Analysis of RIOPA Data. Environ. Sci. Technol. 41, 7315–7321.
https://doi.org/10.1021/es070037k
Mi, D., Fang, H., Zhao, Y., Zhong, L., 2017. Birth weight and type 2 diabetes: A meta-analysis. Exp Ther
Med 14, 5313–5320. https://doi.org/10.3892/etm.2017.5234
Mikati, I., Benson, A.F., Luben, T.J., Sacks, J.D., Richmond-Bryant, J., 2018. Disparities in Distribution
of Particulate Matter Emission Sources by Race and Poverty Status. Am J Public Health 108,
480–485. https://doi.org/10.2105/AJPH.2017.304297
Morello-Frosch, R., Zuk, M., Jerrett, M., Shamasunder, B., Kyle, A.D., 2011. Understanding The
Cumulative Impacts Of Inequalities In Environmental Health: Implications For Policy. Health
Affairs 30, 879–887. https://doi.org/10.1377/hlthaff.2011.0153
Morrison, J.L., Duffield, J.A., Muhlhausler, B.S., Gentili, S., McMillen, I.C., 2010. Fetal growth
restriction, catch-up growth and the early origins of insulin resistance and visceral obesity.
Pediatric Nephrology (Berlin, Germany) 25, 669–677. https://doi.org/10.1007/s00467-009-1407-
3
Pagels, J., Wierzbicka, A., Fors, E., Isaxon, C., Dahl, A., Gudmundsson, A., Swietlicki, E., Bohgard, M.,
2009. Chemical composition and mass emission factors of candle smoke particles. Journal of
Aerosol Science 40, 193–208. https://doi.org/10.1016/j.jaerosci.2008.10.005
Pedersen, M., Giorgis-Allemand, L., Bernard, C., Aguilera, I., Andersen, A.-M.N., Ballester, F., Beelen,
R.M.J., Chatzi, L., Cirach, M., Danileviciute, A., Dedele, A., Eijsden, M. van, Estarlich, M.,
Fernández-Somoano, A., Fernández, M.F., Forastiere, F., Gehring, U., Grazuleviciene, R.,
Gruzieva, O., Heude, B., Hoek, G., Hoogh, K. de, van den Hooven, E.H., Håberg, S.E., Jaddoe,

65
V.W.V., Klümper, C., Korek, M., Krämer, U., Lerchundi, A., Lepeule, J., Nafstad, P., Nystad,
W., Patelarou, E., Porta, D., Postma, D., Raaschou-Nielsen, O., Rudnai, P., Sunyer, J., Stephanou,
E., Sørensen, M., Thiering, E., Tuffnell, D., Varró, M.J., Vrijkotte, T.G.M., Wijga, A., Wilhelm,
M., Wright, J., Nieuwenhuijsen, M.J., Pershagen, G., Brunekreef, B., Kogevinas, M., Slama, R.,
2013. Ambient air pollution and low birthweight: a European cohort study (ESCAPE). The
Lancet Respiratory Medicine 1, 695–704. https://doi.org/10.1016/S2213-2600(13)70192-9
Petersen, R., 2019. Racial and Ethnic Disparities in Adult Obesity in the United States: CDC’s Tracking
to Inform State and Local Action. Prev. Chronic Dis. 16. https://doi.org/10.5888/pcd16.180579
Price, P.N., Shehabi, A., Chan, R.W., Gadgil, A.J., 2006. Indoor-Outdoor Air Leakage of Apartments and
Commercial Buildings (No. LBNL-60682, 1010623). https://doi.org/10.2172/1010623
Rich, D.Q., Liu, K., Zhang, Jinliang, Thurston, S.W., Stevens, T.P., Pan, Y., Kane, C., Weinberger, B.,
Ohman-Strickland, P., Woodruff, T.J., Duan, X., Assibey-Mensah, V., Zhang, Junfeng, 2015.
Differences in Birth Weight Associated with the 2008 Beijing Olympics Air Pollution Reduction:
Results from a Natural Experiment. Environ Health Perspect 123, 880–887.
https://doi.org/10.1289/ehp.1408795
Risnes, K.R., Vatten, L.J., Baker, J.L., Jameson, K., Sovio, U., Kajantie, E., Osler, M., Morley, R.,
Jokela, M., Painter, R.C., Sundh, V., Jacobsen, G.W., Eriksson, J.G., Sørensen, T.I.A., Bracken,
M.B., 2011. Birthweight and mortality in adulthood: a systematic review and meta-analysis. Int J
Epidemiol 40, 647–661. https://doi.org/10.1093/ije/dyq267
Rosofsky, A., Levy, J.I., Breen, M.S., Zanobetti, A., Fabian, M.P., 2019. The impact of air exchange rate
on ambient air pollution exposure and inequalities across all residential parcels in Massachusetts.
J Expo Sci Environ Epidemiol 29, 520–530. https://doi.org/10.1038/s41370-018-0068-3
Rossen, L.M., 2014. Neighbourhood economic deprivation explains racial/ethnic disparities in overweight
and obesity among children and adolescents in the USA. J Epidemiol Community Health 68,
123–129. https://doi.org/10.1136/jech-2012-202245
Rundle, A., Hoepner, L., Hassoun, A., Oberfield, S., Freyer, G., Holmes, D., Reyes, M., Quinn, J.,
Camann, D., Perera, F., Whyatt, R., 2012. Association of childhood obesity with maternal
exposure to ambient air polycyclic aromatic hydrocarbons during pregnancy. American Journal of
Epidemiology 175, 1163–1172. https://doi.org/10.1093/aje/kwr455
Sarnat, J.A., Koutrakis, P., Suh, H.H., 2000. Assessing the relationship between personal particulate and
gaseous exposures of senior citizens living in Baltimore, MD. J Air Waste Manag Assoc 50,
1184–1198. https://doi.org/10.1080/10473289.2000.10464165
Savitz, D.A., Bobb, J.F., Carr, J.L., Clougherty, J.E., Dominici, F., Elston, B., Ito, K., Ross, Z., Yee, M.,
Matte, T.D., 2014. Ambient Fine Particulate Matter, Nitrogen Dioxide, and Term Birth Weight in
New York, New York. Am J Epidemiol 179, 457–466. https://doi.org/10.1093/aje/kwt268
Schembari, A., de Hoogh, K., Pedersen, M., Dadvand, P., Martinez, D., Hoek, G., Petherick, E.S.,
Wright, J., Nieuwenhuijsen, M.J., 2015. Ambient Air Pollution and Newborn Size and Adiposity
at Birth: Differences by Maternal Ethnicity (the Born in Bradford Study Cohort). Environ Health
Perspect 123, 1208–1215. https://doi.org/10.1289/ehp.1408675
Slama, R., Morgenstern, V., Cyrys, J., Zutavern, A., Herbarth, O., Wichmann, H.-E., Heinrich, J., 2007.
Traffic-Related Atmospheric Pollutants Levels during Pregnancy and Offspring’s Term Birth
Weight: A Study Relying on a Land-Use Regression Exposure Model. Environ Health Perspect
115, 1283–1292. https://doi.org/10.1289/ehp.10047
Smith, C.J., Ryckman, K.K., Barnabei, V.M., Howard, B.V., Isasi, C.R., Sarto, G.E., Tom, S.E., Van
Horn, L.V., Wallace, R.B., Robinson, J.G., 2016. The impact of birth weight on cardiovascular
disease risk in the Women’s Health Initiative. Nutr Metab Cardiovasc Dis 26, 239–245.
https://doi.org/10.1016/j.numecd.2015.10.015
Stabile, L., Fuoco, F.C., Buonanno, G., 2012. Characteristics of particles and black carbon emitted by
combustion of incenses, candles and anti-mosquito products. Building and Environment 56, 184–
191. https://doi.org/10.1016/j.buildenv.2012.03.005

66
Stieb, D.M., Chen, L., Eshoul, M., Judek, S., 2012. Ambient air pollution, birth weight and preterm birth:
A systematic review and meta-analysis. Environmental Research 117, 100–111.
https://doi.org/10.1016/j.envres.2012.05.007
Stieb, D.M., Chen, L., Hystad, P., Beckerman, B.S., Jerrett, M., Tjepkema, M., Crouse, D.L., Omariba,
D.W., Peters, P.A., van Donkelaar, A., Martin, R.V., Burnett, R.T., Liu, S., Smith-Doiron, M.,
Dugandzic, R.M., 2016. A national study of the association between traffic-related air pollution
and adverse pregnancy outcomes in Canada, 1999–2008. Environmental Research 148, 513–526.
https://doi.org/10.1016/j.envres.2016.04.025
Suh, H.H., Zanobetti, A., 2010. Exposure Error Masks The Relationship Between Traffic-Related Air
Pollution and Heart Rate Variability (HRV). J Occup Environ Med 52, 685–692.
https://doi.org/10.1097/JOM.0b013e3181e8071f
Sun, X., Luo, X., Zhao, C., Zhang, B., Tao, J., Yang, Z., Ma, W., Liu, T., 2016. The associations between
birth weight and exposure to fine particulate matter (PM2.5) and its chemical constituents during
pregnancy: A meta-analysis. Environmental Pollution 211, 38–47.
https://doi.org/10.1016/j.envpol.2015.12.022
Umer, A., Hamilton, C., Cottrell, L., Giacobbi, P., Innes, K., Kelley, G.A., Neal, W., John, C., Lilly, C.,
2020. Association between birth weight and childhood cardiovascular disease risk factors in West
Virginia. J Dev Orig Health Dis 11, 86–95. https://doi.org/10.1017/S204017441900045X
Upadhyay, R.P., Naik, G., Choudhary, T.S., Chowdhury, R., Taneja, S., Bhandari, N., Martines, J.C.,
Bahl, R., Bhan, M.K., 2019. Cognitive and motor outcomes in children born low birth weight: a
systematic review and meta-analysis of studies from South Asia. BMC Pediatrics 19, 35.
https://doi.org/10.1186/s12887-019-1408-8
Vilanova, C.S., Hirakata, V.N., de Souza Buriol, V.C., Nunes, M., Goldani, M.Z., da Silva, C.H., 2019.
The relationship between the different low birth weight strata of newborns with infant mortality
and the influence of the main health determinants in the extreme south of Brazil. Popul Health
Metr 17, 15. https://doi.org/10.1186/s12963-019-0195-7
Visentin, S., Grumolato, F., Nardelli, G.B., Di Camillo, B., Grisan, E., Cosmi, E., 2014. Early origins of
adult disease: low birth weight and vascular remodeling. Atherosclerosis 237, 391–399.
https://doi.org/10.1016/j.atherosclerosis.2014.09.027
Wahabi, H.A., Alzeidan, R.A., Fayed, A.A., Mandil, A., Al-Shaikh, G., Esmaeil, S.A., 2013. Effects of
secondhand smoke on the birth weight of term infants and the demographic profile of Saudi
exposed women. BMC Public Health 13, 341. https://doi.org/10.1186/1471-2458-13-341
Watkins, W.J., Kotecha, S.J., Kotecha, S., 2016. All-Cause Mortality of Low Birthweight Infants in
Infancy, Childhood, and Adolescence: Population Study of England and Wales. PLoS Med 13.
https://doi.org/10.1371/journal.pmed.1002018
Whitaker, A.H., Feldman, J.F., Lorenz, J.M., Shen, S., McNicholas, F., Nieto, M., McCulloch, D., Pinto-
Martin, J.A., Paneth, N., 2006. Motor and Cognitive Outcomes in Nondisabled Low-Birth-
Weight Adolescents: Early Determinants. Arch Pediatr Adolesc Med 160, 1040.
https://doi.org/10.1001/archpedi.160.10.1040
Wilcox, A.J., Weinberg, C.R., Basso, O., 2011. On the Pitfalls of Adjusting for Gestational Age at Birth.
Am J Epidemiol 174, 1062–1068. https://doi.org/10.1093/aje/kwr230
Wylie, B.J., Kishashu, Y., Matechi, E., Zhou, Z., Coull, B., Abioye, A.I., Dionisio, K.L., Mugusi, F.,
Premji, Z., Fawzi, W., Hauser, R., Ezzati, M., 2017. Maternal exposure to carbon monoxide and
fine particulate matter during pregnancy in an urban Tanzanian cohort. Indoor Air 27, 136–146.
https://doi.org/10.1111/ina.12289
Zeger, S.L., Thomas, D., Dominici, F., Samet, J.M., Schwartz, J., Dockery, D., Cohen, A., 2000.
Exposure measurement error in time-series studies of air pollution: concepts and consequences.
Environ Health Perspect 108, 419–426.
Zhang, Yaqun, Cao, S., Xu, X., Qiu, J., Chen, M., Wang, D., Guan, D., Wang, C., Wang, X., Dong, B.,
Huang, H., Zhao, N., Jin, L., Bai, Y., Duan, X., Liu, Q., Zhang, Yawei, 2016. Metals

67
compositions of indoor PM2.5, health risk assessment, and birth outcomes in Lanzhou, China.
Environ Monit Assess 188, 325. https://doi.org/10.1007/s10661-016-5319-y
Zhu, X., Liu, Y., Chen, Y., Yao, C., Che, Z., Cao, J., 2015. Maternal exposure to fine particulate matter
(PM2.5) and pregnancy outcomes: a meta-analysis. Environ Sci Pollut Res 22, 3383–3396.
https://doi.org/10.1007/s11356-014-3458-7

68
SUPPLEMENTARY MATERIALS
Table S.2.1. Bivariate Analysis of Personal PM 2.5 and Birthweight by Categorical Participant
Characteristics (N = 205).
Personal PM 2.5 Birthweight
Variable

n Mean (SD)
(ug/m
3
)
p-value Mean (SD)
(grams)
p-value
Infant Sex 0.727 0.465
Female 105 21.9 (15.7) 3,266.9 (462.6)
Male 100 22.7 (18.6) 3,316.7 (508.7)
Race 0.485 0.355
Hispanic 166 22.7 (17.1) 3,294.4 (480.7)
Black, Non-Hispanic 23 22.8 (19.9) 3,185.0 (475.5)
Other, Non-Hispanic 16 17.3 (12.9) 3,410.6 (542.2)
Parity 0.243 0.042
Yes 130 23.8 (19.0) 3,338.8 (493.5)
No 69 19.5 (13.1) 3,180.6 (408.3)
Missing 6 21.2 (13.8) 3,531.7 (867.6)
Diabetes 0.252 0.318
Normal 136 20.8 (14.7) 3,249.4 (453.7)
Glucose Intolerant 46 25.4 (21.2) 3,343.6 (514.3)
Gestational Diabetes Mellitus
10 21.0 (5.6) 3,440.5 (628.1)
Chronic Diabetes 13 28.0 (27.1) 3,428.5 (572.1)
Smoking 0.988 0.302
Ever 43 22.3 (11.6) 3,223.2 (512.4)
Never 162 22.3 (18.4) 3,309.2 (477.6)
Pre-pregnancy BMI 0.751 0.271
Normal 62 21.3 (14.7) 3,208.5 (440.1)
Overweight 64 23.6 (18.9) 3,318.3 (521.5)
Obese 79 22.1 (17.5) 3,334.1 (486.1)
Education 0.069 0.472
< 12
th
grade 49 24.2 (19.6) 3,324.9 (523.5)
Completed High School 64 21.6 (13.2) 3,214.8 (396.1)
Some College 62 24.9 (21.0) 3,309.7 (532.2)
Completed College 30 15.4 (7.3) 3,361.0 (494.3)
Household Income 0.025 0.323
$50,000 + 14 33.8 (38.5) 3,480.3 (676.9)
$30,000-$49,999 29 17.3 (10.4) 3,289.0 (485.8)
$15,000-$29.9999 46 21.0 (15.2) 3,341.8 (380.9)
Less than $15,000 41 25.5 (18.9) 3,314.8 (590.8)
Don’t Know 75 21.1 (11.4) 3,212.8 (432.5)
Season 0.072 0.853
Warm 40 19.3 (13.2) 3,308.1 (547.1)
Cool 64 26.3 (16.7) 3,262.9 (516.5)
Transition 101 21.0 (18.4) 3,302.4 (441.0)

69
Questionnaire Variables used in Interaction Analyses
Time Activity Patterns
Time Spent Indoors
How much of the time did you spend indoors (at your residence, or someone else's residence)?
a

0.249 0.005
None and a little of the time 31 25.2 (15.5) 3,065.3 (397.2)
Most and all of the time 173 21.4 (16.7) 3,332.0 (490.6)
Thinking back to a typical weekday in this past week, approximately how many hours (out of 24 hours in
total) did you spend indoors?
b

0.115 0.459
≤ 16 hours 48 25.7 (15.7) 3,245.6 (445.1)
> 16 hours 157 21.3 (17.5) 3,305.1 (497.2)
Time Spent Outdoors
How much of the time did you spend outdoors (not commuting in a car, bus or train)?
a

0.410 0.164
None and a little of the time 130 21.3 (16.8) 3,327.3 (498.5)
Most and all of the time 74 23.3 (16.1) 3,228.6 (460.5)
Thinking back to a typical weekday in this past week, approximately how many hours (out of 24 hours in
total) did you spend outdoors?
b

0.724 0.554
< 8 hours 162 22.1 (18.3) 3,301.6 (492.8)
≥ 8 hours 43 23.1 (12.1) 3,252.1 (458.0)
Home Characteristics and Ventilation
Home Type
Which best describes the home in which you currently live most of the time?
b

0.200 0.383
A single-family home
(No joining walls)
72 18.5 (12.7) 3,286.0 (468.9)
2-4 Units 48 25.4 (21.1) 3,252.4 (428.3)
5+ units 68 23.2 (15.8) 3,269.0 (501.1)
Missing 17 25.8 (24.0) 3,511.2 (612.8)
How much of the time were windows (or porch/balcony doors if applicable) open in your home, when
you were there with the sampler?
a

0.058 0.420
None and a little of the time 81 24.7 (22.4) 3,325.4 (553.9)
Most and all of the time 123 20.2 (11.0) 3,269.1 (436.9)
On average, how much of the time were the windows open in your home this past week?
b

0.015 0.474
None and a little of the time 84 25.8 (23.3) 3,320.4 (526.4)
Most and all of the time 121 19.9 (10.5) 3,270.9 (455.4)
How much of the time was the air conditioner used in your home, when you were there with the
sampler?
a

0.955 0.013
A little, most, and all of the time 53 22.1 (19.8) 3,433.5 (504.6)
None of the time 151 22.0 (15.3) 3,241.6 (471.2)
Do you use air conditioning in your home?
b

0.872 0.076

70
Yes 85 22.1 (17.7) 3,362.5 (505.1)
No 120 22.5 (16.8) 3,240.6 (465.9)
Indoor Sources
Cooking
How much of the time were you close to smoke or fumes from cooking (yourself, or nearby cooking by
someone else) e.g. burnt toast, barbeque, stir fry, etc.?
a

0.884 0.913
A little, most, or all of the time 80 21.8 (17.7) 3,286.8 (453.6)
None of the time 124 22.1 (15.9) 3,294.5 (508.0)
Since we last saw you/spoke to you in your first/second trimester, on average, how many times a week do
you cook (using the stove/range/oven, not microwave)?
b

0.009 0.834
1 or more times a week 192 21.5 (15.5) 3,293.0 (480.4)
Never 13 34.3 (31.5) 3,263.9 (570.9)
Candle or Incense
How much of the time were you close to smoke from candles or incense burning nearby?
a

0.254 0.929
A little, most, or all of the time 50 24.3 (14.1) 3,286.2 (504.1)
None of the time 154 21.2 (17.2) 3,293.2 (482.0)
Notes:
a
From exit questionnaire administered to participants after completing the 48-hour personal exposure monitoring period;
b
From 3
rd
trimester survey; p-values are analysis of variance (ANOVA) tests; bolded = statistically significant at p-value < 0.05;
PM2.5 = particulate matter with aerodynamic diameter less than 2.5µm; BMI = body mass index; SD = standard deviation; g =
grams; kg = kilograms; transition = spring and autumn.

Table S.2.2. Bivariate Analysis of Personal PM 2.5 and Birthweight by Participant
Characteristics (N = 205).
PM 2.5 Birthweight
Variable Correlation p-value Correlation p-value
Birthweight (g) 0.08 0.273

Total Weight Gain (kg) -0.05 0.472 0.29 <0.001
Gestational Age (weeks) -0.08 0.231 0.43 <0.001
Maternal Age (years) 0.02 0.776 0.02 0.738
3
rd
Trimester Average
Temperature (°C)
-0.15 0.038 0.02 0.782
Notes: Pearson’s Correlation Coefficients was used to determine correlations; PM2.5 = particulate matter
with aerodynamic diameter less than 2.5µm; g = grams; kg = kilogram; bolded = statistically significant at
p-value < 0.05.

71
CHAPTER 3
EFFECTS OF IN-UTERO PERSONAL EXPOSURE TO PM 2.5 SOURCES ANDCOMPONENTS
ON BIRTHWEIGHT
ABSTRACT
Background: In-utero exposure to particulate matter with aerodynamic diameter less than 2.5µm (PM 2.5)
and specific sources and components of PM 2.5 have been linked with lower birthweight. However,
previous results have been mixed, likely due to heterogeneity in sources impacting PM 2.5 and its chemical
composition across regions, and due to measurement error from using ambient data to estimate personal
exposure. Therefore, we investigated the effect of PM 2.5 sources and their high-loading components on
birthweight using data from 198 women in the 3
rd
trimester from the MADRES cohort personal PM 2.5
exposure monitoring sub-study.
Methods: The mass contributions of six major sources of personal PM 2.5 exposure were estimated for 198
pregnant women in the 3
rd
trimester along with their 17 high-loading chemical components. Components
were ascertained from optical carbon and X-ray fluorescence analyses of 48-hour integrated personal
PM 2.5 Teflon filter samples, which were used in a prior study to derive the PM 2.5 sources using the EPA
Positive Matrix Factorization (EPA PMF v5.0) model. Single- and multi-pollutant linear regressions were
used to evaluate the association between personal PM 2.5 sources and birthweight, adjusting for key
covariates. Additionally, high-loading components were evaluated with birthweight individually and in
models further adjusted for PM 2.5 mass.
Results: Participants were predominately Hispanic (81%), never smokers (80%), with a mean (SD)
gestational age of 39.1 (1.5) weeks and age of 28.2 (6.0) years. Mean birthweight was 3,295.8g (484.1)
and mean PM 2.5 exposure was 21.3 (14.4) µg/m
3
. A 1 SD increase in the mass contribution of the fresh sea
salt source was statistically significantly associated with a 101.1g decrease in birthweight (95% CI:-201.8,
-0.5), while aged sea salt was marginally significantly associated with lower birthweight (β =-69.9; 95%
CI:-141.3, 1.6). The components Mg, Na, and Cl were significantly associated with lower birthweight,
which remained after adjusting for PM 2.5 mass.

72
Conclusions: Overall, this study found evidence that major sources of personal PM 2.5 including fresh and
aged sea salt that are of outdoor origin were negatively associated with birthweight.

73
INTRODUCTION
Low birthweight (LBW) is an endemic negative health outcome, with an estimated 8.3% of
newborns born in the United States (U.S.) having a birthweight below 2,500 grams (g) (Martin et al., 2018).
This number is higher worldwide with Southern Asia and Sub-Saharan Africa facing the greatest burden
(Blencowe et al., 2019). LBW is known to be associated with several negative health outcomes, including
infant mortality (Vilanova et al., 2019; Watkins et al., 2016), later-life obesity (Jornayvaz et al., 2016),
diabetes (Mi et al., 2017), and cardiovascular disease (Huxley et al., 2007; Umer et al., 2020) and impaired-
cognitive development (Upadhyay et al., 2019; Whitaker et al., 2006). Many of these negative health
outcomes often disproportionately affect race/ethnicity groups, for example, obesity and type-2 diabetes
prevalence are highest in Hispanic and Black communities (Petersen, 2019; Rossen, 2014).
Combined with the greater burden of some negative health outcomes faced by Hispanic and Black
communities, they also experience the greatest cumulative burden of air pollution exposure (Bell & Ebisu,
2012; Mikati et al., 2018). Various epidemiological studies, including several meta-analyses, across the
world have found a modest association between ambient air pollution exposure during the in utero period
and birthweight, including being LBW (Lamichhane et al., 2015; X. Li et al., 2017; Pedersen et al., 2013;
Stieb et al., 2016). Of these ubiquitous air pollutants, a moderate association between particulate matter
(PM) with an aerodynamic diameter less than 2.5µm (PM 2.5) and lower birthweight has been found using
both ambient (Dadvand et al., 2013; Huang et al., 2015; Savitz et al., 2014; Schembari et al., 2015) and
personal monitoring approaches for PM 2.5 exposure assessment (Jedrychowski et al., 2004, 2009). Exposure
to PM 2.5 likely creates a hostile intrauterine environment which is hypothesized to explain its toxic effects,
and while the biological mechanisms behind this effect are still emerging, studies suggest that oxidative
stress, DNA methylation, and endocrine disruption may be at play and may lead to placental inflammation
and growth restriction (Clemente et al., 2016; Z. Li et al., 2019).
However, while reviews have concluded that there is a relationship between exposure to ambient
PM 2.5 and decreases in birthweight (X. Li et al., 2017; Stieb et al., 2016; Sun et al., 2016), there is a great
deal of heterogeneity in the findings, possibly due to measurement error from estimating personal exposure

74
of PM 2.5 of outdoor origin from ambient data (Carroll, 2005; Kioumourtzoglou et al., 2014; Zeger et al.,
2000). Additionally, this may also be explained in part by the fact that PM 2.5 is actually a mixture of organic
and inorganic chemicals, with specific sources and/or components PM 2.5 being shown to differ in their
toxicity with respect to birthweight when investigating the effects of PM 2.5 constituents measured at ambient
monitoring stations (Bell et al., 2010; Bell & Ebisu, 2012; Darrow et al., 2011; Laurent et al., 2016a; Ng et
al., 2017).
In Study 1 of this dissertation, no association was found between total personal exposure to PM 2.5
and birthweight using personal monitoring in the third trimester; however, specific indoor sources (such as,
candle and incense smoke) and exposure to PM 2.5 more impacted by sources of outdoor origin appeared to
be more strongly associated with lower birthweight. These earlier findings are in line with the literature,
with studies showing exposure to traffic-related sources including on-road gasoline and diesel traffic were
negative associated with birthweight (Basu et al., 2014; Laurent et al., 2016a; Wilhelm & Ritz, 2003). Also,
secondhand smoke (SHS) exposure is associated with reduced birthweight and increased risk of LBW
(Khader et al., 2011; Wahabi et al., 2013). To date it is unclear whether risk associated with these sources
may be driven by exposure to the source itself as a unique mixture of pollutants or by any of its specific
marker or signature chemical components. Therefore, in this analysis, we propose to investigate the effects
of both major contributing sources of personal PM 2.5 and their high-loading or signature components on
birthweight.
This is an important question since the chemical composition of PM 2.5 in conjunction with other
properties like size distribution determine particles’ toxicity. Chemical composition of PM 2.5 exposures may
also differ by race/ethnicity with researchers finding that Hispanic individuals are exposed to elevated levels
of 13 out of 14 PM 2.5 components they investigated compared to non-Hispanic Whites in California (Basu
et al., 2014). Of these components, several were linked to increased risk of LBW and reduced birthweight
(Basu et al., 2014; Sun et al., 2016). For example, Basu et al. (2014) found statistically significant reductions
in birthweight per 1 IQR increase in exposure to vanadium (𝛽:-32; 95% CI: -38, -27), titanium (𝛽: -15; 95%
CI: -17,-13), zinc (𝛽:-10; 95% CI: -12, -7), and elemental carbon (𝛽:-16; 95% CI: -19, -14) (Basu et al.,

75
2014). Furthermore, a meta-analysis from Sun et al (2016), corroborated these findings, but they also found
that other components including silicon and nickel were elevated in Hispanic neighborhoods compared to
Non-Hispanic White neighborhoods and were also negatively associated with birthweight.
However, as with the existing health literature on the effects of PM 2.5 mass, there is a great deal of
heterogeneity of results when looking at the mixture of sources and components that compose it (Ebisu et
al., 2014; Ebisu & Bell, 2012; Laurent et al., 2014; Sun et al., 2016). One possibility is that individual
exposure is assigned using estimates of outdoor concentrations at the residential level (Harris et al., 2014),
which fail to account for time-activity patterns and infiltration of outdoor pollutants indoors (Gray et al.,
2010; Kioumourtzoglou et al., 2014) which introduces exposure measurement error. However, this error
might be exacerbated when investigating the effects of chemical components of PM 2.5 because of their
greater spatial variability relative to PM 2.5 mass concentration as a whole (Bell et al., 2011). Additionally,
total personal exposure to major contributing sources and components of PM 2.5 also includes contributions
from indoor sources and from personal activity or behavior related sources, and not just outdoor sources.
Therefore, in this analysis we aimed to investigate the relationship between exposure to six
chemically derived major sources of personal PM 2.5 in the 3
rd
trimester of pregnancy with infant
birthweight. To accomplish this goal, we leveraged personal measurements of exposure to PM 2.5 mass and
its components and source apportionment models. We fit single and multi-pollutant models for the
sources and also investigated the independent effects of their high-loading “signature” chemical
components in an environmental health disparities population.
MATERIALS AND METHODS
Study Population
This work takes place in a 214-participant personal PM 2.5 exposure monitoring sub-study nested
within the Maternal and Developmental Risks from Environmental and Social Stressors (MADRES) study,
an ongoing prospective cohort study of ~900 pregnant, primarily Hispanic, low-income pregnant women
in Los Angeles County (Bastain et al., 2019). The aim of MADRES is to investigate the cumulative impact

76
of environmental pollutants and psychosocial, behavioral, and built environmental risk factors on maternal
and infant health outcomes as described in more detail elsewhere (Bastain et al., 2019). Briefly, pregnant
women were enrolled into the cohort through partnerships with four prenatal care clinics in Los Angeles,
CA from November 2015. Eligibility for this study included: 1) at least 18 years old, 2) a singleton birth,
3) less than 30-weeks’ gestation at recruitment, 4) HIV negative, 5) having no physical, mental, or cognitive
disability that would prevent the participant from providing informed consent, and 5) not currently
incarcerated.
Participants were recruited by trained, bilingual MADRES staff members during a 3
rd
trimester
visit to the University of Southern California (USC) clinic, where consenting women were asked to
participate in the in-utero personal exposure monitoring sub-study for a 48-hour monitoring period. This
sample was comparable to the larger MADRES cohort on key demographics, birth outcomes, and ambient
air pollution metrics.
Personal PM 2.5 Exposure Monitoring
Total personal PM 2.5 exposure was measured over an integrated 48-hours monitoring period in the
3
rd
trimester using a custom designed sampling protocol between October 2016 and February 2020.
Participants were shown and provided with instructions by trained staff members on correct use of the
personal exposure monitoring device, which was housed in a crossbody purse. Instructions included a
demonstration of how to wear the purse, making sure to keep the sampling inlet located on the purse
shoulder strap free from obstructions and in the breathing zone. Additionally, participants were requested
to wear the device as much as possible during normal daily activities, with a limited number of exceptions,
including, driving, showering, sleeping, etc. Participants were asked to keep the sampling device safe and
away from water, high humidity (such as showering), heat, children, and pets, and when unable to wear the
device, place it as near as possible, such as on the passenger seat if driving, and a side-table while sleeping.
The purse contained a Gilian Plus Datalogging Pump (Sensidyne Inc.) connected to a Harvard
PM 2.5 Personal Environmental Monitor (PEM) with a pre-weighed 37mm Pall Teflo filter. The device was

77
programmed to start at midnight the day after enrollment into the sub-study, and actively sampled at a 50%
cycle and a 1.8 liters per minute (LPM) flow rate. The sampling device was programmed to shut off after
the 48-hour sampling period and collected by staff members the following day when a brief exit survey was
conducted. The devices were then transferred to the USC Exposure Analytics lab for analysis. Pump data
was downloaded, evaluated for errors, and stored securely. Filters was equilibrated within a dedicated
chamber and gravimetrically weighed in temperature and relative humidity-controlled glove box using an
MT-5 microbalance (Mettler Toledo, Inc.) to obtain PM 2.5 mass concentration reported in μg/m
3
.
Personal PM 2.5 Sources
Six major contributing sources of personal PM 2.5 were used in this analysis, obtained from an earlier
source apportionment analysis of these personal exposure filter samples using the EPA Positive Matrix
Factorization model (EPA PMF v5.0) by my colleague Dr. Yan Xu. The PMF analysis used PM 2.5 mass
and the concentrations of 36 components (33 elements and 3 optical carbon fractions) as inputs to derive
the six sources and their predicted mass contributions. The elements and carbon species were obtained from
X-ray fluorescence (XRF) and multiwavelength optical absorption carbon speciation analyses, respectively,
at the Research Triangle Institute International, Inc (described in more detail below). For the purposes of
this current study, only 17 (14 elements and 3 optical carbon fractions) high-loading components or
signature tracers of the six sources were investigated with birthweight including:
1) Traffic (BC, Zn, Ba; 2.6%), 2) Secondhand smoke (BrC, ETS, Br; 65.3%), 3) Aged Sea Salt (S,
Na, Mg; 4.3%), 4) Fresh Sea Salt (Cl, Na, Mg; 4.7%), Fuel oil (Cu, Ni, V; 11.7%), and Crustal (Si, Ca, Ti,
Al; 11.5%). Although vanadium (V) had S/N < 0.2, it was included (and set to “weak”) in the PMF analysis
because it was considered important tracer elements for traffic and fuel oil sources.
The PMF model (Paatero, 1997) has been used extensively in earlier air pollution source
apportionment studies (Bell et al., 2010; Habre et al., 2020; Ottone et al., 2020) Briefly, the PMF model
solves the following equation:
𝑥
!"
=∑ 𝑔
!#
∗𝑓
"#
+𝑒
!"
$
#%&
eq. (1);

78
where 𝑥
!"
is the concentration of species j in sample i, 𝑔
!#
is the mass contribution of source k to sample i,
𝑓
"#
is the loading of species j on source k, and 𝑒
!"
is the residual error for sample i and species j. Eq.(1) is
solved by minimizing Q (object function) through an iterative process:
𝑄 = ∑ ∑ .
'
!"
(
!"
/
)
*
"%&
+
!%&
eq. (2);
where 𝑢
!"
is the uncertainty of species j in sample i as provided by the laboratory from the speciation
analysis. Using the measurement uncertainty value for each sample, elements included in the analysis are
filtered based on a PMF-calculated signal-to-noise ratio (S/N) which generally corresponds to the
proportion of samples that are measured confidently above the limit of detection. Based on this S/N ratio,
elements are determined to be bad (S/N≤0.2) and excluded from the analysis; whereas elements that are
considered weak (0.2≤S/N≤2) are included but down-weighted and elements that are considered strong
are retained (S/N>2). Based on the analyst’s expert determination of optimal number of factors/sources,
the “fingerprints” or source profiles (f matrix in eq(1)) and the mass contributions (g matrix in eq(1)) of the
major sources contributing to the variation in PM 2.5 measurements are calculated. Both the source profiles
(chemical signatures) and the variation in the source mass contributions (estimated quantities of PM 2.5
coming from each source) were examined to identify and label these sources.
Elemental Speciation Analysis Using X-ray Fluorescence
Elemental data was obtained via an X-ray fluorescence analysis that determined the daily elemental
composition of the personal PM 2.5 mass from the loaded 37mm Pall Teflo filters and have been described
elsewhere (Gill, 2014). Briefly, this is a non-destructive process that allowed for the quantification of
specific elements captured onto the filters. Each element concentration measurement also comes with an
uncertainty value, which was used in the PMF source apportionment to down-weight samples with greater
analytical noise in terms of their influence on the solution (Paatero, 1997). Concentrations of elements
(reported in ng/m
3
) identified in the source apportionment analysis as markers or high-loading species in
the source profiles were used in this current analysis. These included: aluminum (Al), barium (Ba), bromine

79
(Br), calcium (Ca), chlorine (Cl), copper (Cu), magnesium (Mg), nickel (Ni), silicon (Si), sodium (Na),
sulfur (S), titanium (Ti), vanadium (V), and zinc (Zn).
Optical Carbon Fractions Analysis
A multiwavelength optical absorption approach was used to measure concentrations of several
carbon fractions (reported in μg/m
3
) in the personal PM 2.5 samples, including: 1) Black Carbon (BC), 2)
Brown Carbon (BrC), and 3) Environmental Tobacco Smoke (ETS). This method is described in more
detail elsewhere (Lawless, 2004), and its performance is consistent with other carbon apportionment
approaches (Yan et al., 2011). Briefly, this method uses an integrating sphere radiometer which measures
the difference in absorption of transmitted light passed through a mass loaded Teflo filter. Each of the three
carbon components measured with this approach have a different optical density at varying wavelengths,
which allows for the identification and quantification of their respective concentration from their optical
properties. For the purposes of this study, ETS refers to the carbon fraction concentration obtained via this
multiwavelength optical analysis, while the secondhand smoke (SHS) source (explained above) refers to
one of the six major contributing sources of personal PM 2.5 identified in the PMF analysis. This source had
high loadings of several different but highly correlated components, namely ETS and BrC.
Birthweight Outcome
Infant birthweight (grams) was abstracted from participants’ electronic medical record (EMR).
Given that birthweight and gestational age are closely linked, birthweight-for-gestational age z-scores that
were either sex or parity specific were also assessed, as described in (Aris et al., 2019). However, the results
were not materially different from continuous birth; therefore, only the continuous birthweight models are
presented.
Covariate Data
Possible covariates were determined a priori from the air pollution and birth outcomes literature.
Factors assessed included maternal demographics, pregnancy and birth outcomes, study design
characteristics, and meteorology. Self-report data were collected during the MADRES cohort follow-up

80
through a sequence of staff administered in-person and telephone-based questionnaires. All questionnaires
were available in either English or Spanish. Anthropometric assessments were conducted via regular clinic
visits. Due to the timing of this personal monitoring study in the 3
rd
trimester of pregnancy, data primarily
came from the 3
rd
trimester visit, the personal monitoring study exit survey, and birth-related datasets and
variables, with additional variables assessed at the baseline visit including race/ethnicity and pre-pregnancy
body mass index (BMI; kg/m
2
).
Additional pregnancy and birth-related covariates were also evaluated for confounding. Gestational
age at birth (GA; weeks) was estimated with a hierarchical approach of differing methods from the preferred
ultrasound measurement of crown-rump length at <14 weeks gestation (60%), ultrasound measurement of
fetal biparietal diameter at < 28 weeks’ gestation (30%), and from physicians’ clinical estimate from EMR
(10%). Parity was defined as 1 or more previous births and included a missing category with 6 participants
so as not to lose sample size. Infant sex was obtained through electronic medical records (EMR).
Maternal demographics included the following: Age at baseline (continuous; years), education
(completed <12th grade, completed high school, at least some college), household income (less than
$15,000, $15,000 – 29,999, $30,000+, don’t know), and diabetes status (no diabetes, glucose intolerant,
diabetes (chronic and gestational)). Race/ethnicity was recategorized from the NIH categories into
Hispanic, non-Hispanic Black, and non-Hispanic Other due the variable distribution to save degrees of
freedom in our regression analysis. Pre-pregnancy BMI (continuous; kg/m
2
) was calculated from self-
reported pre-pregnancy weight and standing height measured by MADRES staff at participants’ first visit
by either a stadiometer (Perspectives Enterprises model PE-AIM-101) or EMR. Self-report weight was
chosen because participants entered the study at differing weeks gestation.
Meteorological factors evaluated in this study included ambient air temperature (Celsius)
(calculated as average of minimum and maximum air temperature) and relative humidity (%), averaged
over the 3
rd
trimester and estimated at the residential location based on a high-resolution (4km x 4km)
gridded surface meteorological dataset (Abatzoglou, 2011).

81
Statistical Analysis
Descriptive Statistics
Descriptive statistics of key sample demographics and mean and standard deviations were
calculated for concentrations of personal PM 2.5 mass, six PMF-derived sources of personal PM 2.5, and 17
high-loading components. The distribution of birthweight, personal PM 2.5 mass concentration, and each
source and component were investigated to identify any issues with normality and potential influential
points. Bivariate analyses using Kruskal-Wallis one-way analysis of variance tests and Spearman’s
correlation coefficients were conducted between personal PM 2.5 mass, its major contributing sources, and
birthweight by various sample characteristics to elicit any additional potential confounders for our
regression analysis, in addition to those identified in previous literature.
Linear Regression Models
Single- and multi-pollutant linear regression models were used to investigate the association
between in-utero exposure to major personal PM 2.5 sources and birthweight, adjusting for key covariates.
The effect of total personal PM 2.5 on birthweight was included in relevant tables for comparison purposes.
PM 2.5 sources that were not highly correlated with one another, as determined by a bivariate Spearman
correlation analysis and/or a variance inflation factor (VIF) below 10 in the regression, were used to
evaluate the effect of each source on birthweight, adjusting for one or more other PM 2.5 sources.
Additionally, PM 2.5 has been shown to differ by the sex of the infant, therefore, this study evaluated whether
the effect of each PM 2.5 source exposure on birthweight was modified by sex. Non-linear effects were
evaluated by modeling each PM 2.5 source on birthweight using generalized additive models (GAMs) and
assessing Akaike information criterion (AICs) vs. linear regression models. As a sensitivity analysis, the
association of each PM 2.5 source on birthweight was examined only among full-term births (37 weeks or
older gestation) to assess whether the pre-term births impacted the associations seen in the full sample.
Next, to evaluate whether it is the PM 2.5 source or any of its high-loading components that is driving
the observed association between sources and birthweight, the effect of the 17 high-loading PM 2.5

82
components on birthweight was investigated. Additionally, because PM 2.5 mass concentration may be
related to both birthweight and the concentration of the PM 2.5 components (especially more abundant ones),
further analyses adjusting component models for PM 2.5 mass were performed via two approaches. The first
was by adjusting for PM 2.5 mass in the component models by directly including it as a simple covariate or
potential confounder. The second approach was to create component residuals by regressing the mass
concentration of each PM 2.5 component (dependent) on the total PM 2.5 mass concentration (independent).
The residuals were then used in birthweight models as exposures, thereby eliminating the influence of
variable PM 2.5 mass concentration on component models in health analyses when used as exposures in our
health analyses (Mostofsky et al., 2012).
Due to concerns with outliers being influential as determined by model diagnostics in 3 out of 6
main source models, a multivariate K-nearest neighbor outlier detection analysis was conducted in JMP
Pro 16 (SAS Institute, Inc., Cary, NC, USA). This was used to identify outliers up to a distance of 8 nearest
neighbors along the concentrations of all six personal PM 2.5 sources. This analysis allowed us to objectively
identify data points that were materially different than the overall sample across six dimensions. All effect
estimates and 95% confidence intervals were scaled and reported per 1 SD increase in the main exposure
of interest (shown in Table 3.1.). An alpha of 0.05 was selected as a priori significance level for our main
exposure/outcome analyses, while 0.10 was used for the infant sex interaction analyses. Model diagnostics
were conducted to ensure models were not affected by multi-collinearity or influential points. The analysis
was conducted using SAS v9.4 (SAS Institute, Inc., Cary, NC, USA).
RESULTS
Descriptive Statistics
Of the 214 participants in the personal exposure monitoring study, nine were removed due to
incomplete or erroneous personal PM 2.5 mass exposure data or birth outcome data. Four participants did not
have PM 2.5 source data and were removed from the analysis. The multivariate k-nearest neighbor outlier
detection analysis revealed three outliers in terms of personal exposure to the six sources. These were
excluded from further analysis. However, given these points were very influential in the models, results

83
including and excluding them are presented side-by-side for completeness in this analysis. This resulted in
a sample of 198 mother-infant dyads used in the final models (201 in the outlier included models).
Sample characteristics for the full sample are presented in Table 3.1. The participants of this study
were predominantly Hispanic (81%), lower income (45% had income below $30,000), with a mean age of
28 years, and 63% having had a previous pregnancy. Around 70% of the women were overweight or obese,
and 34% had glucose intolerance or diabetes (chronic or gestational). Participants’ infants were 51% female
and had a mean (SD) birthweight of 3,295.8 (484.1) grams and gestational age of 39.1 (1.5) weeks at time
of birth.
Table 3.1. Sample Participant Characteristics (N = 201).
Variable
Mean (SD)
or n (%) Variable
Mean (SD)
or n (%)
Birthweight (g) 3,295.8 (484.1)
Pre-pregnancy BMI
(kg/m
2
) 28.9 (6.8)
Maternal age (years) 28.2 (6.0) Normal 61 (30.4%)
Gestational Age (weeks) 39.1 (1.5) Overweight 61 (30.4%)
Sex Obese 79 (39.3%)
Female 103 (51.2%) Parity
Male 98 (48.8%) No 68 (33.8%)
Race/ethnicity Yes 127 (63.2%)
Hispanic 163 (81.1%) Missing 6 (3.0%)
Black, Non-Hispanic 22 (11.0%) Maternal Income
Other, Non-Hispanic 16 (8.0%) Less than $15,000 41 (20.4%)
Education $15,000-$29,999 45 (22.4%)
<12th grade 48 (23.9%) $30,000+ 42 (20.9%)
Completed High
School 65 (32.3%) Don't know 73 (36.3%)
Some college+ 88 (43.8%) Smoking History
Diabetes Never 160 (79.6%)
Normal 133 (66.2%) Ever 41 (20.4%)
Glucose Intolerant 45 (22.4%) Outdoor Temperature (°C) 19.02 (3.5)
Diabetes^ 23 (11.4%)

Notes: BMI = body mass index; ^ diabetes = chronic or gestational diabetes.

Total personal PM 2.5 exposure had a mean (SD) of 21.3 (14.4) μg/m
3
. The average mass
contributions of the six personal PM 2.5 sources were as follows: SHS 11.8 (9.3) μg/m
3
, crustal 2.3 (3.9)
μg/m
3
, fuel oil 2.1 (1.6) μg/m
3
, aged sea salt 0.9 (0.9) μg/m
3
, fresh sea salt 0.8 (2.1) μg/m
3
, and traffic 0.5

84
(0.6) μg/m
3
(Table 3.2.). Carbon fractions concentrations were generally similar; however, ETS was the
highest with mean of 1.4 μg/m
3
but also had the highest variability with a SD of 5.5 μg/m
3
.

Sulfur, sodium,
silicon, and chlorine were the elements measured at highest concentrations, with mean (SD) of 397.7
(283.9), 311.1 (305.7), 165.7 (200.0), and 129.5 (259.4) ng/m
3
, respectively (Table 3.2.).
Table 3.2. Summary Statistics of PM 2.5 Sources and Components Concentrations (N = 201).
Pollutant Mean SD Pollutant Mean SD
Personal PM 2.5 mass (μg/m
3
) 21.3 14.4 Elements (ng/m
3
)
Aluminum (Al) 12.3 48.0
Sources (μg/m
3
) Barium (Ba) 14.6 13.7
Secondhand Smoke (SHS) 11.8 9.3 Bromine (Br) 3.0 3.2
Crustal 2.3 3.9 Calcium (Ca) 85.8 143.2
Fuel Oil 2.1 1.6 Chlorine (Cl) 129.5 259.4
Aged Sea Salt 0.9 0.9 Copper (Cu) 18.7 12.2
Fresh Sea Salt 0.8 2.1 Magnesium (Mg) 40.0 63.9
Traffic 0.5 0.6 Nickel (Ni) 2.3 2.7

Sodium (Na) 311.1 305.7
Optical Carbon Fractions (μg/m
3
) Silicon (Si) 165.7 200.0
Black Carbon (BC) 1.0 1.5 Sulfur (S) 397.7 283.9
Brown Carbon (BrC) 1.1 0.8 Titanium (Ti) 10.0 11.9
Environmental Tobacco Smoke (ETS) 1.4 5.5 Vanadium (V) 0.6 1.2
Zinc (Zn) 13.2 16.9
Note: PM2.5 = particulate matter with an aerodynamic diabetes less than 2.5µm.

Sociodemographic and Other Characteristics in Relation to Sources

The relationships between personal PM 2.5 mass and its six sources with key demographics are
presented in Table S.3.1. of the supplement. There was no significant difference in personal PM 2.5 mass
exposure by sociodemographic and other covariates. However, personal PM 2.5 mass concentration was
highest in Hispanic (21.6 μg/m
3
) and non-Hispanic Black (22.0 μg/m
3
) participants compared to non-
Hispanic Others (17.4 μg/m
3
). Additionally, participants with less than 12
th
grade of education had the
highest personal PM 2.5 concentration (23.8 μg/m
3
), followed by those who completed high school (21.6
μg/m
3
) and at least some college (19.8 μg/m
3
). Diabetics (chronic and gestational) were exposed to roughly
2 and 4 μg/m
3
greater PM 2.5 that those with glucose intolerance and non-diabetes, respectively. There was
no discernable difference in personal PM 2.5 exposure by smoking history status or infant sex.

85
As for personal PM 2.5 sources, traffic was positively correlated with gestational age (Spearman ρ =
0.19; p = 0.006) and was significantly higher in Hispanic participants (mean 0.5 μg/m
3
) compared to non-
Hispanic Black (0.4 μg/m
3
) and non-Hispanic Other (0.2 μg/m
3
; p = 0.031). There were significant
differences in SHS exposure by diabetes status, with glucose intolerant participants having the highest
concentrations (15.2 μg/m
3
), compared to 12.2 μg/m
3
for diabetics, and 10.6 μg/m
3
for non-diabetes (p =
0.016). There was a marginally significant difference in SHS exposure by maternal income (< $15,000:
13.3 μg/m
3
, $15,000-$29,999: 10.5 μg/m
3
, $30,000+: 9.6 μg/m
3
, don’t know: 13.2 μg/m
3
;

p = 0.058). Aged
sea salt was significantly positively correlated with outdoor temperature (ρ = 0.65; p < 0.001). While not
statistically significant, participants with less than 12
th
grade education had higher exposure to the crustal
source (3.3 μg/m
3
), compared to 2.1 μg/m
3
and 1.8 μg/m
3
for those who completed high school or had some
college, respectively.
Correlation of Personal PM 2.5 Sources and Components
Traffic was significantly positively correlated with crustal (ρ = 0.32; p < 0.001) and significantly
negatively correlated with fresh sea salt (ρ = -0.22; p = 0.002) and fuel oil (ρ = -0.17; p = 0.017) (Table
3.3). SHS was significantly negatively correlated with aged sea salt (ρ = -0.24; p < 0.001), and fresh sea
salt (ρ = -0.26; p < 0.001). Crustal was moderately correlated with fuel oil (ρ = 0.15; p = 0.034). A
Spearman correlation matrix for the components is presented in Table S.3.2. Spearman correlations
varied between -0.58 for ETS and BC (p < 0.001) to 0.86 for Na and Mg (p < 0.001). The majority of
correlations were positive. Other extremely high correlations included elements that loaded on the same
source, such as Si and Ca (ρ = 0.76; p < 0.001), ETS and BrC (ρ = 0.66; p < 0.001), and Ti and Ca (ρ =
0.66; p < 0.001).

86
Table 3.3. Spearman’s Correlation Coefficients for Major Contributing Sources of Personal PM 2.5 (N =
201).
Variables Traffic
Secondhan
d Smoking
Aged Sea
Salt
Fresh Sea
Salt
Fuel Oil Crustal
Traffic 1.00

Secondhand
Smoke
-0.09 1.00

Aged Sea Salt -0.04 -0.24 1.00

Fresh Sea Salt -0.22 -0.26 0.09 1.00

Fuel Oil -0.17 -0.02 -0.08 0.00 1.00

Crustal 0.32 -0.01 -0.10 -0.09 0.15 1.00
Notes: PM2.5 = particulate matter with an aerodynamic diabetes less than 2.5µm; correlations colored from negative (red)
to positive (blue);
bolded = p-value < 0.05.

Associations of Personal PM 2.5 Sources and Components with Birthweight

All linear regression models included the following covariates: sex of infant, GA, maternal age,
race/ethnicity, parity, diabetes status, maternal education, smoking status (never/ever), and temperature.
Even though this study assessed SHS as a source of PM 2.5, it did not correlate strongly with our smoking
covariate (never/ever smoker). However, this smoking covariate did seem to be a confounder and impact
our main effects, therefore, it was kept within the model. Multi-pollutant models were conducted with up
to four personal PM 2.5 sources included in each model; however, three- and four-pollutant model results did
not materially differ. Therefore, only single- and two-pollutant models are shown (Table 3.4.). Overall, in
the fully adjusted final models, personal PM 2.5 exposure was not associated with birthweight (β = -33.5;
95% CI: -103.2, 36.1). There was an average significant decrease of 101.1g (95% CI: -201.8, -0.5) in
birthweight per 1 SD increase in the fresh sea salt source. This result remained after adjusting individually
for fuel oil and crustal sources but became marginally significant after adjustment for aged sea salt. A 1 SD
increase in aged sea salt was marginally associated with a 69.9g (95% CI: -141.3, 1.6) decrease in

87
birthweight. This remained after further adjustment for fresh sea salt, traffic, fuel oil, and crustal sources.
There was no association with birthweight for traffic, SHS, fuel oil, and crustal sources.
Table 3.4. Single- and Two-pollutant Associations Between PM 2.5 Sources and Birthweight.
Final Model (N = 198) Outliers Included (N = 201)
Model β 95% CI β 95% CI
Single-pollutant models
Personal PM 2.5 mass
-33.5 -103.2 36.1 42.1 -15.6 99.9
Personal PM 2.5 Sources

Traffic
22.4 -35.7 80.4 25.3 -34.6 85.3
Secondhand Smoke (SHS)
-12.1 -69.5 45.4 -19.5 -78.3 39.3
Aged Sea Salt
-69.9 -141.3 1.6 -45.8 -115.0 23.3
Fresh Sea Salt
-101.1 -201.8 -0.5 16.3 -42.5 75.1
Fuel Oil
16.5 -39.8 72.7 25.7 -32.1 83.5
Crustal
-13.8 -101.1 73.5 72.8 15.6 130.0
Two-pollutant models
Traffic
Adjusted for SHS
21.6 -36.8 79.9 24.1 -36.1 84.3
Adjusted for Aged Sea Salt
20.1 -37.6 77.7 22.9 -37.1 82.8
Secondhand Smoking

Adjusted for Fuel Oil
-11.7 -69.3 45.9 -19.0 -77.9 39.8
Adjusted for Traffic
-10.4 -68.1 47.3 -17.9 -76.9 41.1
Adjusted for Crustal
-12.1 -69.7 45.5 -19.6 -77.6 38.4
Aged Sea Salt

Adjusted for Fresh Sea Salt
-68.9 -139.79 2.04 -57.05 -129.7 15.6
Adjusted for Traffic
-68.8 -140.43 2.82 -44.12 -113.5 25.3
Adjusted for Fuel Oil
70.0 -141.6 1.6 -45.5 -114.7 23.7
Adjusted for Crustal
-69.9 -141.6 1.7 -41.2 -109.5 27.1
Fresh Sea Salt

Adjusted for Aged Sea Salt
-99.8 -199.7 0.2 31.1 -30.5 92.7
Adjusted for Fuel Oil
-102.4 -203.3 -1.5 15.9 -43.0 74.8
Adjusted for Crustal
-105.6 -207.5 -3.8 20.0 -38.0 78.1
Fuel Oil

Adjusted for Aged Sea Salt
16.8 -39.1 72.6 25.3 -32.4 83.0
Adjusted for Fresh Sea Salt
18.6 -37.2 74.4 25.5 -32.4 83.4
Adjusted for SHS
16.2 -40.2 72.6 25.4 -32.5 83.3
Crustal

Adjusted for Aged Sea Salt
-14.3 -100.9 72.4 70.9 13.7 128.1
Adjusted for SHS
-13.9 -101.4 73.6 72.8 15.5 130.1
Adjusted for Fresh Sea Salt
-26.9 -114.4 60.6 73.8 16.4 131.1
Notes: PM2.5 = particulate matter with an aerodynamic diabetes less than 2.5µm; SHS = secondhand smoke PM2.5 source; β =
change in birthweight per 1 SD increase in pollutant; CI = confidence intervals; all models were adjusted for gestational age at
birth, maternal age, race/ethnicity, infant sex, parity, diabetes status, temperature, maternal education, and personal smoking
history.

88
Two high-loading components of fresh sea salt and aged sea salt, Na (β = -89.4; 95% CI: -163.0, -
15.7) and Mg (β = -153.2; 95% CI: -248.7, -57.7), were statistically associated with lower birthweight
(Figure 3.1.). Neither secondhand smoking (SHS) as a source nor its components were associated with
birthweight despite negative relationships. The effect estimates did not materially change after adjustment
for personal PM 2.5 mass as a covariate or when looking at PM 2.5 component residuals. In a sensitivity
analysis in just full-term births, the results did not materially change (Table S.3.3.).
Figure 3.1. Associations Between High-Loading Components of the Six Personal PM 2.5 Sources and
Birthweight in Single-Pollutant Models, Adjusting for Personal PM 2.5 Mass Concentration, and Using
the Component Residuals as the Main Exposure, Respectively.

Notes: PM2.5 = particulate matter with an aerodynamic diabetes less than 2.5µm; significance < 0.05; β = change in
birthweight per 1 SD increase in pollutant; adj = adjusting; all models were adjusted for gestational age at birth, maternal age,
race/ethnicity, infant sex, parity, diabetes status, temperature, maternal education, and personal smoking history.

Effects of Personal PM 2.5 Sources on Birthweight by Infant Sex
While not meeting statistical significance, the effect of personal PM 2.5 on birthweight was more
negative in males (β = -63.3g; 95% CI: -169.4, 42.8) compared to females (β = -11.6g; 95% CI: -13.0, 79.9;
Table 3.5.). The effect of the crustal source on birthweight was significantly modified by infant sex, with
males observing a -86.5g (95% CI: -191.8, 18.7) decrease in birthweight per 1 SD increase in exposure to

89
the crustal source compared to a 127.3g (95% CI: -18.2, 272.8) increase for females (Interaction: p = 0.019).
Additionally, the effect of the fuel oil source on birthweight was -31.1 (95% CI: -114.8, 52.6) for males and
55.5 (95% CI: -20.2, 131.3) for females, which was marginally significant (Interaction: p = 0.133).
Table 3.5. Estimated Change in Birthweight (g) per 1 SD Increase in Pollutant by Infant Sex (N = 198).
Variable β 95% CI Interaction p-value
Interaction of Personal PM 2.5 Mass by Infant
Sex 0.464
Female -11.6 -103.0 79.9

Male -63.3 -169.4 42.8

Interaction of Traffic by Infant Sex

0.777
Female 7.2 -113.6 128.0

Male 26.7 -38.90 92.4

Interaction of SHS by Infant Sex

0.991
Female -12.3 -85.4 60.8

Male -11.7 -102.2 78.9

Interaction of Aged Sea Salt by Infant Sex

0.605
Female -82.4 -168.4 3.7

Male -51.5 -151.4 48.4

Interaction of Fresh Sea Salt by Infant Sex

0.522
Female -141.2 -300.3 18.0

Male -75.0 -203.8 53.8

Interaction of Fuel Oil by Infant Sex

0.133
Female 55.5 -20.3 131.3

Male -31.1 -114.8 52.6

Interaction of crustal by Infant Sex

0.019
Female 127.3 -18.2 272.8
Male -86.5 -191.8 18.7
Notes: All models are adjusted for gestational age at birth, maternal age, race/ethnicity, infant sex, parity, diabetes status,
temperature, maternal education, and personal smoking history; SHS = secondhand smoke source.
DISCUSSION
This analysis investigated the effect of major contributing sources of personal PM 2.5 exposure on
birthweight in the 3
rd
trimester within the MADRES in-utero personal exposure monitoring study. To our
knowledge, this is the first analysis relating chemically derived sources of personal PM 2.5 exposure with
birthweight in a largely lower income, Hispanic, health disparities population leveraging “gold standard”
personal exposure monitoring data. Prior studies reported concerns about exposure misclassification due to

90
spatial variability of PM 2.5 components when using ambient monitoring data to assign individual exposure
(Bell et al., 2011). This present study used sources derived from personal exposure monitoring data, which
may remove much of this concern by measuring PM 2.5 components for each individual in their personal
breathing zone (Gray et al., 2010).
Overall, this study found that total personal PM 2.5 was not statistically significantly associated with
birthweight; however, some of its more specific sources, including fresh sea salt and aged sea salt were
negatively associated with birthweight. Both sources contained Na and Mg as high-loading components,
which were the elements that were most negatively associated with birthweight in this analysis. The
findings in this present study differ to prior studies which found no association between sea or marine salt
sources and reduced birthweight (Bell et al., 2010; Wilhelm et al., 2012). There are several possible
explanations for the different findings. For example, Bell et al. (2010) was conducted in Connecticut and
Massachusetts, while the present study was in California, likely resulting in regional differences in PM 2.5
sources and components with differing toxicities. Also, this current study used personal exposure
monitoring, while prior studies used ambient approaches, which may have led to attenuated or biased health
effects. Prior studies looking at the effects of outdoor Na concentrations on birthweight are limited;
however, Basu et al. (2014) found that Na was associated with decreased birthweight in California.
Similarly, outdoor concentration of sodium ion (Na
+
) was associated with lower birthweight on the East
Coast (Ebisu et al., 2014). The relationship between Mg has been less studied, but a study in California
found a no association between Mg and birthweight (Laurent et al., 2016b). It is not apparent why sea salt,
particularly fresh sea salt, has the most adverse effect of birthweight in our study. It is possible fresh sea
salt may be correlated with offshore marine shipping emissions given they both originate over the ocean
and may be transported to receptor locations under similar meteorological and wind conditions.
Components of marine shipping emissions (Ni, V, and EC) have been associated with lower birthweight
(Bell et al., 2014; Laurent et al., 2014). However, in our study, the fuel oil personal exposure source (with
loadings of Ni and V) was poorly correlated with fresh sea salt. Fresh sea salt as an outdoor source of PM 2.5
is also correlated with being closer to the coast and the ports in Los Angeles, CA (Habre et al., 2021);

91
however, this spatial pattern was not apparent in our personal monitoring study for the fresh sea salt or for
the fuel oil sources. It is possible that the fuel oil personal exposure source in our study is capturing impacts
of heavy-duty machinery and industrial equipment that burn heavier residual fuel oil, which is very common
in Los Angeles, CA rather than picking up shipping emission signals from the ports (Arhami et al., 2008).
Fresh and aged sea salt did differ in an important way, with fresh sea salt containing higher loadings
of Cl which is replaced with S as fresh sea salt undergoes photochemical reactions and becomes aged sea
salt (Habre et al., 2020; Hasheminassab et al., 2014). While the two sources share Na and Mg as high-
loading components, they are not highly correlated (Table 3.3.). Differences in their impacts on birthweight
may be due to other components or factors that correlate with them. For example, there was a -86.17g (95%
CI: -172.70, 0.36) decrease in birthweight per a 1 SD increase in Cl, compared to a -24.29g (95%CI: -97.87,
49.29) decrease for S. Previously, S was associated with decreased birthweight (Basu et al., 2014; Pedersen
Marie et al., 2016), while results for Cl have been mixed with two studies finding statistically significant
reductions in birthweight (Basu et al., 2014; Ebisu et al., 2014), while another found no association with
LBW risk (Ebisu & Bell, 2012). Interestingly, these studies found an association with Cl and decreased
birthweight in New England and California, highlighting that this may not be a local phenomenon.
Other considerations for why aged sea salt might be negatively associated with birthweight is
through secondary formation processes. Aged sea salt as a personal exposure source was highly correlated
with outdoor ozone and temperature in our study. Both ozone and aged sea salt undergo chemical aging
and transformation processes in the atmosphere under similar conditions of high temperature which could
explain this correlation (Gard et al., 1998; Habre et al., 2021; Hasheminassab et al., 2014). However, 8-
hour maximum ozone concentration was not significantly associated with birthweight for the same 48 hours
sampling period or the whole third trimester within our sample (estimated at the residence using inverse
distance weighted squared spatial interpolation, data not shown). This suggests there may be other
processes, factors, or co-exposures associated with personal exposure to fresh and aged sea salt – both of
clear outdoor origin – that may also be negatively associated with birthweight.

92
When components are more negatively associated than the source itself, it may imply that that
particular compound is more toxic than the whole mixture. Efforts to identify sources or important
components can lead to actionable interventions. For example, researchers found that stricter caps in Europe
on sulfur content in marine fuel led to 22% reductions in sulfur dioxide gas and 6% reductions in PM 2.5,
which in turn resulted in a 7% reduced risk of LBW (Lindgren, 2021). In California, while there has been
a general reduction in ambient PM 2.5 concentrations over the past 20+ years due to regulatory interventions
(Lurmann et al., 2015), this study adds to the literature that specific sources or components of PM 2.5 still
place pregnant mothers at risk of adverse birth outcomes.
This study found that SHS and two high-loading components (ETS and BrC) were negatively but
not significantly associated with reduced birthweight; however, the effect estimates for ETS and BrC were
more negative that the SHS source itself. For example, 1 SD increase in ETS and BrC was associated with
a -35.2g (95%CI: -94.3, 23.8) and -31.8g (95%CI: -89.9, 26.4) reduction in birthweight, respectively,
compared to -12.1g (95%CI: -69.5, 45.4) for the SHS source. Overall, our results agree with prior studies
that found a negative association between SHS and birthweight (Hawsawi et al., 2015; Wahabi et al., 2013).
Additionally, a prior study investigating the effect of total personal PM 2.5 on birthweight, found that self-
reported pre-natal SHS was negatively but not-significantly associated with reduced birthweight
(Jedrychowski et al., 2009).
This current study did not find an association between the personal traffic exposure source or its
components. This differs to prior studies that have generally concluded that traffic-related exposures are
related to lower birthweight (Dong et al., 2022) and an increased risk of LBW (Laurent et al., 2014, 2016b;
Wilhelm et al., 2012). Personal exposure to zinc (Zn) was not associated with lower birthweight, which is
inconsistent with the consensus of prior studies (Basu et al., 2014; Sun et al., 2016). Different spatial scales
may explain the differences observed between studies as prior literature has used ambient monitoring
compared to personal monitoring in this study. Similarly, personal exposure to the fuel oil source was not
associated with reductions in birthweight in this study; however, there was evidence that infant sex may

93
modify this relationship with a -31.1g (95% CI: -114.84 52.6) decrease in birthweight for males compared
to a 55.5g (95% CI: -20.3, 131.3) increase in birthweight for females.
This study found a small non-significant negative association between crustal exposure and lower
birthweight. Additionally, a statistically significant interaction was found by infant sex, with negative
effects seen in males and positive effects in females. Together with fuel oil, this provides evidence for
potential differences in the underlying biomechanisms of how air pollution affects health, which may
interact with sex-based biological differences among fetuses. Notably, unlike other components we
investigated, effect estimates for crustal components were the most altered when adjusted for personal PM 2.5
mass. This suggests that the effect of these components may be the result of the effect of PM 2.5 mass because
PM 2.5 may be correlated with both the component and the health outcome (Mostofsky et al., 2012), possibly
because crustal components are more abundant and therefore more correlated with PM 2.5 mass.
This study has several strengths, including the use of chemical speciation data and source
apportionment derived PM 2.5 sources, from “gold standard” personal exposure monitoring data. Prior
studies that assessed PM 2.5 sources and components (Laurent et al., 2016a; Sun et al., 2016) used outdoor
measurements which do not account for exposures that occur indoors or in-transit due to time-activity
patterns, indoor sources, and infiltration of outdoor sources into the home. MADRES is a well characterized
cohort, with a vast array of individual level covariate data available, making this an excellent study for this
research question to be conducted. Furthermore, this study provided evidence for the effect of PM 2.5 sources
on birthweight in a health disparities population, which may experience not just greater burden of adverse
health outcomes and environment exposures, but also lower access to health care and resources to alleviate
the impact of such burdens (Mahajan et al., 2021; Wang et al., 2013).
The sample size of this study is a potential limitation with a final working sample of 198 (201 in
the outlier included model), which while small for population-based health studies, it is actually fairly large
for personal exposure monitoring studies (Dadvand et al., 2012; Sarnat et al., 2000; Suh & Zanobetti, 2010).
However, even with this potential limitation, we were able to detect several associations between major
sources of PM 2.5 and their respective high-loading components. Another possible limitation is from

94
participation bias due to differences in the type of expecting mothers that chose to participate in the personal
exposure monitoring study component of the MADRES cohort study. We did not observe any material
differences between participants who chose to participate in the personal exposure monitoring study
compared to the larger cohort, expect that they were slightly more likely to have had a prior child (data not
shown).
Finally, this study was conducted over a 48-hour sampling period in the 3
rd
trimester, which may
not be representative of typical or longer duration. However, the correlation between personal PM 2.5 and
ambient PM 2.5 for the same 48-hour sampling period was very similar to that with the 3
rd
trimester average
ambient PM 2.5 (Study 1). Also, we found reasonable concordance between time-activity patterns two
different questionnaire sources, including an exit survey after the 48-hour sampling period, and a 3
rd

trimester questionnaire (data not shown), suggesting these are not drastically different. Together, these
increase confidence in the representativeness of the measurements in terms of what participants are truly
experiencing considering time-activity patterns and outdoor exposures are major contributing factors to
personal PM 2.5.
CONCLUSION
Overall, this study found evidence that major outdoor sources of personal PM 2.5 including fresh sea
salt, aged sea salt, and to a lesser extent, sources of more indoor and human behavior related origin like
SHS and crustal, were negatively associated with birthweight in a health disparities population in Los
Angeles, CA. Additionally, the effect of crustal and fuel oil sources differed by infant sex with negative
associations seen in boys compared to positive associations for girls.

95
REFERENCES

Aris, I. M., Kleinman, K. P., Belfort, M. B., Kaimal, A., & Oken, E. (2019). A 2017 US Reference for
Singleton Birth Weight Percentiles Using Obstetric Estimates of Gestation. Pediatrics, 144(1),
e20190076. https://doi.org/10.1542/peds.2019-0076
Basu, R., Harris, M., Sie, L., Malig, B., Broadwin, R., & Green, R. (2014). Effects of fine particulate
matter and its constituents on low birth weight among full-term infants in California.
Environmental Research, 128, 42–51. https://doi.org/10.1016/j.envres.2013.10.008
Bell, M. L., Belanger, K., Ebisu, K., Gent, J. F., Lee, H. J., Koutrakis, P., & Leaderer, B. P. (2010).
Prenatal Exposure to Fine Particulate Matter and Birth Weight. Epidemiology (Cambridge,
Mass.), 21(6), 884–891. https://doi.org/10.1097/EDE.0b013e3181f2f405
Bell, M. L., & Ebisu, K. (2012). Environmental Inequality in Exposures to Airborne Particulate Matter
Components in the United States. Environmental Health Perspectives, 120(12), 1699–1704.
https://doi.org/10.1289/ehp.1205201
Bell, M. L., Ebisu, K., & Peng, R. D. (2011). Community-level spatial heterogeneity of chemical
constituent levels of fine particulates and implications for epidemiological research. Journal of
Exposure Science & Environmental Epidemiology, 21(4), 372–384.
https://doi.org/10.1038/jes.2010.24
Blencowe, H., Krasevec, J., Onis, M. de, Black, R. E., An, X., Stevens, G. A., Borghi, E., Hayashi, C.,
Estevez, D., Cegolon, L., Shiekh, S., Hardy, V. P., Lawn, J. E., & Cousens, S. (2019). National,
regional, and worldwide estimates of low birthweight in 2015, with trends from 2000: A
systematic analysis. The Lancet Global Health, 7(7), e849–e860. https://doi.org/10.1016/S2214-
109X(18)30565-5
Carroll, R. J. (2005). Measurement Error in Epidemiologic Studies. In Encyclopedia of Biostatistics.
American Cancer Society. https://doi.org/10.1002/0470011815.b2a03082
Clemente, D. B. P., Casas, M., Vilahur, N., Begiristain, H., Bustamante, M., Carsin, A.-E., Fernández, M.
F., Fierens, F., Gyselaers, W., Iñiguez, C., Janssen, B. G., Lefebvre, W., Llop, S., Olea, N.,
Pedersen, M., Pieters, N., Santa Marina, L., Souto, A., Tardón, A., … Nawrot, T. S. (2016).
Prenatal Ambient Air Pollution, Placental Mitochondrial DNA Content, and Birth Weight in the
INMA (Spain) and ENVIRONAGE (Belgium) Birth Cohorts. Environmental Health
Perspectives, 124(5), 659–665. https://doi.org/10.1289/ehp.1408981
Dadvand, P., de, N. A., Triguero, -Mas Margarita, Schembari, A., Cirach, M., Amoly, E., Figueras, F.,
Basaga, ña X., Ostro, B., & Nieuwenhuijsen, M. (2012). Surrounding Greenness and Exposure to
Air Pollution During Pregnancy: An Analysis of Personal Monitoring Data. Environmental
Health Perspectives, 120(9), 1286–1290. https://doi.org/10.1289/ehp.1104609
Dadvand, P., Parker, J., Bell, M. L., Bonzini, M., Brauer, M., Darrow, L. A., Gehring, U., Glinianaia, S.
V., Gouveia, N., Ha, E., Leem, J. H., van den Hooven, E. H., Jalaludin, B., Jesdale, B. M.,
Lepeule, J., Morello-Frosch, R., Morgan, G. G., Pesatori, A. C., Pierik, F. H., … Woodruff, T. J.
(2013). Maternal Exposure to Particulate Air Pollution and Term Birth Weight: A Multi-Country
Evaluation of Effect and Heterogeneity. Environmental Health Perspectives, 121(3), 267–373.
https://doi.org/10.1289/ehp.1205575
Darrow, L. A., Klein, M., Strickland, M. J., Mulholland, J. A., & Tolbert, P. E. (2011). Ambient Air
Pollution and Birth Weight in Full-Term Infants in Atlanta, 1994–2004. Environmental Health
Perspectives, 119(5), 731–737. https://doi.org/10.1289/ehp.1002785
Dong, S., Abu-Awad, Y., Kosheleva, A., Fong, K. C., Koutrakis, P., & Schwartz, J. D. (2022). Maternal
exposure to black carbon and nitrogen dioxide during pregnancy and birth weight: Using
machine-learning methods to achieve balance in inverse-probability weights. Environmental
Research, 211, 112978. https://doi.org/10.1016/j.envres.2022.112978

96
Ebisu, K., Belanger, K., & Bell, M. L. (2014). The Association between Airborne PM2.5 Chemical
Constituents and Birth Weight–Implication of Buffer Exposure Assignment. Environmental
Research Letters : ERL [Web Site], 9(8). https://doi.org/10.1088/1748-9326/9/8/084007
Ebisu, K., & Bell, M. L. (2012). Airborne PM2.5 Chemical Components and Low Birth Weight in the
Northeastern and Mid-Atlantic Regions of the United States. Environmental Health Perspectives,
120(12), 1746–1752. https://doi.org/10.1289/ehp.1104763
Gard, E. E., Kleeman, M. J., Gross, D. S., Hughes, L. S., Allen, J. O., Morrical, B. D., Fergenson, D. P.,
Dienes, T., E. Gälli, M., Johnson, R. J., Cass, G. R., & Prather, K. A. (1998). Direct Observation
of Heterogeneous Chemistry in the Atmosphere. Science, 279(5354), 1184–1187.
https://doi.org/10.1126/science.279.5354.1184
Gill, R. (2014). Modern Analytical Geochemistry: An Introduction to Quantitative Chemical Analysis
Techniques for Earth, Environmental and Materials Scientists. Routledge.
https://doi.org/10.4324/9781315844381
Gray, S. C., Edwards, S. E., & Miranda, M. L. (2010). Assessing Exposure Metrics for PM and
Birthweight Models. Journal of Exposure Science & Environmental Epidemiology, 20(5), 469–
477. https://doi.org/10.1038/jes.2009.52
Habre, R., Girguis, M., Urman, R., Fruin, S., Lurmann, F., Shafer, M., Gorski, P., Franklin, M.,
McConnell, R., Avol, E., & Gilliland, F. (2020). Contribution of Tailpipe and Non-tailpipe
Traffic Sources to Quasi-Ultrafine, Fine and Coarse Particulate Matter in Southern California.
Journal of the Air & Waste Management Association, 0(ja), null.
https://doi.org/10.1080/10962247.2020.1826366
Habre, R., Girguis, M., Urman, R., Fruin, S., Lurmann, F., Shafer, M., Gorski, P., Franklin, M.,
McConnell, R., Avol, E., & Gilliland, F. (2021). Contribution of tailpipe and non-tailpipe traffic
sources to quasi-ultrafine, fine and coarse particulate matter in southern California. Journal of the
Air & Waste Management Association, 71(2), 209–230.
https://doi.org/10.1080/10962247.2020.1826366
Harris, G., Thompson, W. D., Fitzgerald, E., & Wartenberg, D. (2014). The association of PM2.5 with
full term low birth weight at different spatial scales. Environmental Research, 134, 427–434.
https://doi.org/10.1016/j.envres.2014.05.034
Hasheminassab, S., Daher, N., Saffari, A., Wang, D., Ostro, B. D., & Sioutas, C. (2014). Spatial and
temporal variability of sources of ambient fine particulate matter
(PM<sub>2.5</sub>) in California. Atmospheric Chemistry and Physics, 14(22),
12085–12097. https://doi.org/10.5194/acp-14-12085-2014
Hawsawi, A. M., Bryant, L. O., & Goodfellow, L. T. (2015). Association Between Exposure to
Secondhand Smoke During Pregnancy and Low Birthweight: A Narrative Review. Respiratory
Care, 60(1), 135–140. https://doi.org/10.4187/respcare.02798
Huang, C., Nichols, C., Liu, Y., Zhang, Y., Liu, X., Gao, S., Li, Z., & Ren, A. (2015). Ambient air
pollution and adverse birth outcomes: A natural experiment study. Population Health Metrics, 13.
https://doi.org/10.1186/s12963-015-0050-4
Huxley, R., Owen, C. G., Whincup, P. H., Cook, D. G., Rich-Edwards, J., Smith, G. D., & Collins, R.
(2007). Is birth weight a risk factor for ischemic heart disease in later life? The American Journal
of Clinical Nutrition, 85(5), 1244–1250. https://doi.org/10.1093/ajcn/85.5.1244
Jedrychowski, W., Bendkowska, I., Flak, E., Penar, A., Jacek, R., Kaim, I., Spengler, J. D., Camann, D.,
& Perera, F. P. (2004). Estimated Risk for Altered Fetal Growth Resulting from Exposure to Fine
Particles during Pregnancy: An Epidemiologic Prospective Cohort Study in Poland.
Environmental Health Perspectives, 112(14), 1398–1402. https://doi.org/10.1289/ehp.7065
Jedrychowski, W., Perera, F., Mrozek-Budzyn, D., Mroz, E., Flak, E., Spengler, J. D., Edwards, S., Jacek,
R., Kaim, I., & Skolicki, Z. (2009). Gender differences in fetal growth of newborns exposed
prenatally to airborne fine particulate matter. Environmental Research, 109(4), 447–456.
https://doi.org/10.1016/j.envres.2009.01.009

97
Jornayvaz, F. R., Vollenweider, P., Bochud, M., Mooser, V., Waeber, G., & Marques-Vidal, P. (2016).
Low birth weight leads to obesity, diabetes and increased leptin levels in adults: The CoLaus
study. Cardiovascular Diabetology, 15(1), 73. https://doi.org/10.1186/s12933-016-0389-2
Khader, Y. S., Al-Akour, N., Alzubi, I. M., & Lataifeh, I. (2011). The association between second hand
smoke and low birth weight and preterm delivery. Maternal and Child Health Journal, 15(4),
453–459. https://doi.org/10.1007/s10995-010-0599-2
Kioumourtzoglou, M.-A., Spiegelman, D., Szpiro, A. A., Sheppard, L., Kaufman, J. D., Yanosky, J. D.,
Williams, R., Laden, F., Hong, B., & Suh, H. (2014). Exposure measurement error in PM2.5
health effects studies: A pooled analysis of eight personal exposure validation studies.
Environmental Health, 13, 2. https://doi.org/10.1186/1476-069X-13-2
Lamichhane, D. K., Leem, J.-H., Lee, J.-Y., & Kim, H.-C. (2015). A meta-analysis of exposure to
particulate matter and adverse birth outcomes. Environmental Health and Toxicology, 30.
https://doi.org/10.5620/eht.e2015011
Laurent, O., Hu, J., Li, L., Cockburn, M., Escobedo, L., Kleeman, M. J., & Wu, J. (2014). Sources and
contents of air pollution affecting term low birth weight in Los Angeles County, California,
2001–2008. Environmental Research, 134, 488–495. https://doi.org/10.1016/j.envres.2014.05.003
Laurent, O., Hu, J., Li, L., Kleeman, M. J., Bartell, S. M., Cockburn, M., Escobedo, L., & Wu, J. (2016a).
Low birth weight and air pollution in California: Which sources and components drive the risk?
Environment International, 92–93, 471–477. https://doi.org/10.1016/j.envint.2016.04.034
Laurent, O., Hu, J., Li, L., Kleeman, M. J., Bartell, S. M., Cockburn, M., Escobedo, L., & Wu, J. (2016b).
A Statewide Nested Case–Control Study of Preterm Birth and Air Pollution by Source and
Composition: California, 2001–2008. Environmental Health Perspectives, 124(9), 1479–1486.
https://doi.org/10.1289/ehp.1510133
Lawless, P. (2004). Multiwavelength absorbance of filter deposits for determination of environmental
tobacco smoke and black carbon. Atmospheric Environment, 38(21), 3373–3383.
https://doi.org/10.1016/j.atmosenv.2004.03.038
Li, X., Huang, S., Jiao, A., Yang, X., Yun, J., Wang, Y., Xue, X., Chu, Y., Liu, F., Liu, Y., Ren, M.,
Chen, X., Li, N., Lu, Y., Mao, Z., Tian, L., & Xiang, H. (2017). Association between ambient
fine particulate matter and preterm birth or term low birth weight: An updated systematic review
and meta-analysis. Environmental Pollution, 227, 596–605.
https://doi.org/10.1016/j.envpol.2017.03.055
Li, Z., Tang, Y., Song, X., Lazar, L., Li, Z., & Zhao, J. (2019). Impact of ambient PM2.5 on adverse birth
outcome and potential molecular mechanism. Ecotoxicology and Environmental Safety, 169,
248–254. https://doi.org/10.1016/j.ecoenv.2018.10.109
Lindgren, S. (2021). The coast is clear: Shipping emission standards, air quality and infant health.
Transportation Research Part D: Transport and Environment, 100, 103067.
https://doi.org/10.1016/j.trd.2021.103067
Lurmann, F., Avol, E., & Gilliland, F. (2015). Emissions Reduction Policies and Recent Trends in
Southern California’s Ambient Air Quality. Journal of the Air & Waste Management Association
(1995), 65(3), 324–335. https://doi.org/10.1080/10962247.2014.991856
Mahajan, S., Caraballo, C., Lu, Y., Valero-Elizondo, J., Massey, D., Annapureddy, A. R., Roy, B., Riley,
C., Murugiah, K., Onuma, O., Nunez-Smith, M., Forman, H. P., Nasir, K., Herrin, J., &
Krumholz, H. M. (2021). Trends in Differences in Health Status and Health Care Access and
Affordability by Race and Ethnicity in the United States, 1999-2018. JAMA, 326(7), 637–648.
https://doi.org/10.1001/jama.2021.9907
Martin, J. A., Hamilton, B. E., Osterman, M. J. K., & Driscoll, A. K. (2018). National Vital Statistics
Reports Volume 68, Number 13, November 30, 2019, Births: Final Data for 2018. 47.
Mi, D., Fang, H., Zhao, Y., & Zhong, L. (2017). Birth weight and type 2 diabetes: A meta-analysis.
Experimental and Therapeutic Medicine, 14(6), 5313–5320.
https://doi.org/10.3892/etm.2017.5234

98
Mikati, I., Benson, A. F., Luben, T. J., Sacks, J. D., & Richmond-Bryant, J. (2018). Disparities in
Distribution of Particulate Matter Emission Sources by Race and Poverty Status. American
Journal of Public Health, 108(4), 480–485. https://doi.org/10.2105/AJPH.2017.304297
Mostofsky, E., Schwartz, J., Coull, B. A., Koutrakis, P., Wellenius, G. A., Suh, H. H., Gold, D. R., &
Mittleman, M. A. (2012). Modeling the Association Between Particle Constituents of Air
Pollution and Health Outcomes. American Journal of Epidemiology, 176(4), 317–326.
https://doi.org/10.1093/aje/kws018
Ng, C., Malig, B., Hasheminassab, S., Sioutas, C., Basu, R., & Ebisu, K. (2017). Source apportionment of
fine particulate matter and risk of term low birth weight in California: Exploring modification by
region and maternal characteristics. Science of The Total Environment, 605–606, 647–654.
https://doi.org/10.1016/j.scitotenv.2017.06.053
Ottone, M., Broccoli, S., Parmagnani, F., Giannini, S., Scotto, F., Bonvicini, L., Luberto, F., Bacco, D.,
Trentini, A., Poluzzi, V., Angelini, P., Colacci, A., Giorgi Rossi, P., & Ranzi, A. (2020). Source-
related components of fine particulate matter and risk of adverse birth outcomes in Northern Italy.
Environmental Research, 186, 109564. https://doi.org/10.1016/j.envres.2020.109564
Paatero, P. (1997). Least squares formulation of robust non-negative factor analysis. Chemometrics and
Intelligent Laboratory Systems, 37(1), 23–35. https://doi.org/10.1016/S0169-7439(96)00044-5
Pedersen, M., Giorgis-Allemand, L., Bernard, C., Aguilera, I., Andersen, A.-M. N., Ballester, F., Beelen,
R. M. J., Chatzi, L., Cirach, M., Danileviciute, A., Dedele, A., Eijsden, M. van, Estarlich, M.,
Fernández-Somoano, A., Fernández, M. F., Forastiere, F., Gehring, U., Grazuleviciene, R.,
Gruzieva, O., … Slama, R. (2013). Ambient air pollution and low birthweight: A European
cohort study (ESCAPE). The Lancet Respiratory Medicine, 1(9), 695–704.
https://doi.org/10.1016/S2213-2600(13)70192-9
Pedersen Marie, Gehring Ulrike, Beelen Rob, Wang Meng, Giorgis-Allemand Lise, Andersen Anne-
Marie Nybo, Basagaña Xavier, Bernard Claire, Cirach Marta, Forastiere Francesco, de Hoogh
Kees, null null, Gruzieva Olena, Hoek Gerard, Jedynska Aleksandra, Klümper Claudia, Kooter
Ingeborg M., Krämer Ursula, Kukkonen Jaakko, … Slama Rémy. (2016). Elemental Constituents
of Particulate Matter and Newborn’s Size in Eight European Cohorts. Environmental Health
Perspectives, 124(1), 141–150. https://doi.org/10.1289/ehp.1409546
Petersen, R. (2019). Racial and Ethnic Disparities in Adult Obesity in the United States: CDC’s Tracking
to Inform State and Local Action. Preventing Chronic Disease, 16.
https://doi.org/10.5888/pcd16.180579
Rossen, L. M. (2014). Neighbourhood economic deprivation explains racial/ethnic disparities in
overweight and obesity among children and adolescents in the USA. Journal of Epidemiology
and Community Health, 68(2), 123–129. https://doi.org/10.1136/jech-2012-202245
Sarnat, J. A., Koutrakis, P., & Suh, H. H. (2000). Assessing the relationship between personal particulate
and gaseous exposures of senior citizens living in Baltimore, MD. Journal of the Air & Waste
Management Association (1995), 50(7), 1184–1198.
https://doi.org/10.1080/10473289.2000.10464165
Savitz, D. A., Bobb, J. F., Carr, J. L., Clougherty, J. E., Dominici, F., Elston, B., Ito, K., Ross, Z., Yee,
M., & Matte, T. D. (2014). Ambient Fine Particulate Matter, Nitrogen Dioxide, and Term Birth
Weight in New York, New York. American Journal of Epidemiology, 179(4), 457–466.
https://doi.org/10.1093/aje/kwt268
Schembari, A., de Hoogh, K., Pedersen, M., Dadvand, P., Martinez, D., Hoek, G., Petherick, E. S.,
Wright, J., & Nieuwenhuijsen, M. J. (2015). Ambient Air Pollution and Newborn Size and
Adiposity at Birth: Differences by Maternal Ethnicity (the Born in Bradford Study Cohort).
Environmental Health Perspectives, 123(11), 1208–1215. https://doi.org/10.1289/ehp.1408675
Stieb, D. M., Chen, L., Hystad, P., Beckerman, B. S., Jerrett, M., Tjepkema, M., Crouse, D. L., Omariba,
D. W., Peters, P. A., van Donkelaar, A., Martin, R. V., Burnett, R. T., Liu, S., Smith-Doiron, M.,
& Dugandzic, R. M. (2016). A national study of the association between traffic-related air

99
pollution and adverse pregnancy outcomes in Canada, 1999–2008. Environmental Research, 148,
513–526. https://doi.org/10.1016/j.envres.2016.04.025
Suh, H. H., & Zanobetti, A. (2010). Exposure Error Masks The Relationship Between Traffic-Related Air
Pollution and Heart Rate Variability (HRV). Journal of Occupational and Environmental
Medicine / American College of Occupational and Environmental Medicine, 52(7), 685–692.
https://doi.org/10.1097/JOM.0b013e3181e8071f
Sun, X., Luo, X., Zhao, C., Zhang, B., Tao, J., Yang, Z., Ma, W., & Liu, T. (2016). The associations
between birth weight and exposure to fine particulate matter (PM2.5) and its chemical
constituents during pregnancy: A meta-analysis. Environmental Pollution, 211, 38–47.
https://doi.org/10.1016/j.envpol.2015.12.022
Umer, A., Hamilton, C., Cottrell, L., Giacobbi, P., Innes, K., Kelley, G. A., Neal, W., John, C., & Lilly,
C. (2020). Association between birth weight and childhood cardiovascular disease risk factors in
West Virginia. Journal of Developmental Origins of Health and Disease, 11(1), 86–95.
https://doi.org/10.1017/S204017441900045X
Upadhyay, R. P., Naik, G., Choudhary, T. S., Chowdhury, R., Taneja, S., Bhandari, N., Martines, J. C.,
Bahl, R., & Bhan, M. K. (2019). Cognitive and motor outcomes in children born low birth
weight: A systematic review and meta-analysis of studies from South Asia. BMC Pediatrics,
19(1), 35. https://doi.org/10.1186/s12887-019-1408-8
Vilanova, C. S., Hirakata, V. N., de Souza Buriol, V. C., Nunes, M., Goldani, M. Z., & da Silva, C. H.
(2019). The relationship between the different low birth weight strata of newborns with infant
mortality and the influence of the main health determinants in the extreme south of Brazil.
Population Health Metrics, 17(1), 15. https://doi.org/10.1186/s12963-019-0195-7
Wahabi, H. A., Alzeidan, R. A., Fayed, A. A., Mandil, A., Al-Shaikh, G., & Esmaeil, S. A. (2013).
Effects of secondhand smoke on the birth weight of term infants and the demographic profile of
Saudi exposed women. BMC Public Health, 13, 341. https://doi.org/10.1186/1471-2458-13-341
Wang, T.-F., Shi, L., Nie, X., & Zhu, J. (2013). Race/Ethnicity, insurance, income and access to care: The
influence of health status. International Journal for Equity in Health, 12(1), 29.
https://doi.org/10.1186/1475-9276-12-29
Watkins, W. J., Kotecha, S. J., & Kotecha, S. (2016). All-Cause Mortality of Low Birthweight Infants in
Infancy, Childhood, and Adolescence: Population Study of England and Wales. PLoS Medicine,
13(5). https://doi.org/10.1371/journal.pmed.1002018
Weisskopf, M. G., & Webster, T. F. (2017). Trade-offs of personal vs. More proxy exposure measures in
environmental epidemiology. Epidemiology (Cambridge, Mass.), 28(5), 635–643.
https://doi.org/10.1097/EDE.0000000000000686
Whitaker, A. H., Feldman, J. F., Lorenz, J. M., Shen, S., McNicholas, F., Nieto, M., McCulloch, D.,
Pinto-Martin, J. A., & Paneth, N. (2006). Motor and Cognitive Outcomes in Nondisabled Low-
Birth-Weight Adolescents: Early Determinants. Archives of Pediatrics & Adolescent Medicine,
160(10), 1040. https://doi.org/10.1001/archpedi.160.10.1040
Wilhelm, M., Ghosh, J. K., Su, J., Cockburn, M., Jerrett, M., & Ritz, B. (2012). Traffic-Related Air
Toxics and Term Low Birth Weight in Los Angeles County, California. Environmental Health
Perspectives, 120(1), 132–138. https://doi.org/10.1289/ehp.1103408
Wilhelm, M., & Ritz, B. (2003). Residential proximity to traffic and adverse birth outcomes in Los
Angeles county, California, 1994-1996. Environmental Health Perspectives, 111(2), 207–216.
Yan, B., Kennedy, D., Miller, R. L., Cowin, J. P., Jung, K., Perzanowski, M., Balletta, M., Perera, F. P.,
Kinney, P. L., & Chillrud, S. N. (2011). Validating a nondestructive optical method for
apportioning colored particulate matter into black carbon and additional components.
Atmospheric Environment (Oxford, England : 1994), 45(39), 7478–7486.
https://doi.org/10.1016/j.atmosenv.2011.01.044
Zeger, S. L., Thomas, D., Dominici, F., Samet, J. M., Schwartz, J., Dockery, D., & Cohen, A. (2000).
Exposure measurement error in time-series studies of air pollution: Concepts and consequences.
Environmental Health Perspectives, 108(5), 419–426.

100
SUPPLEMENTARY MATERIALS
Table S.3.1. Personal PM 2.5 Sources by Key Sample Demographics (N = 201).

Birthweight

PM 2.5
Mass
Traffic SHS
Aged
Sea
Salt
Fresh
Sea
Salt
Fuel
Oil
Crustal

Spearman Correlation Coefficient
Maternal age (years) 0.03 0.02 -0.10 -0.02 0.00 0.01 -0.04 0.10
Gestational Age (weeks) 0.43* -0.06 0.19* -0.09 -0.02 -0.02 0.05 0.08
Temperature (°C) 0.01 -0.12 -0.09 -0.07 0.65* 0.03 -0.12 -0.11
Mean (SD)
b
Sex
Female 3276.9 21.1 0.4 12.1 0.9 0.7 2.0 2.2
Male 3315.6 21.6 0.5 11.5 0.9 1.0 2.08 2.4
Race/ethnicity

*
Hispanic 3293.8 21.6 0.5 12.0 1.0 0.9 2.1 2.3
Black, Non-Hispanic 3226.6 22.0 0.4 11.2 0.7 0.8 2.2 2.4
Other, Non-Hispanic 3410.6 17.4 0.2 10.6 0.5 0.5 1.9 1.3
Education

<12th grade 3346.9 23.8 0.51 11.46 0.83 0.58 1.78 3.30
Completed High School 3208.6 21.6 0.4 12.5 1.0 1.0 2.2 2.1
Some college+ 3332.3 19.8 0.5 11.6 0.8 0.8 2.1 1.8
Diabetes

*
Normal 3257.2 20.2 0.5 10.6 0.9 0.7 2.0 2.3
Glucose Intolerant 3339.2 22.9 0.4 15.2 0.9 1.2 2.1 1.7
Diabetes^ 3433.7 25.0 0.4 12.2 0.7 1.1 2.2 2.8
Pre-pregnancy BMI
(kg/m
2
)

Normal 3208.1 20.2 0.5 11.6 0.9 0.7 2.0 2.3
Overweight 3333.9 21.4 0.4 12.6 0.9 0.9 1.8 2.0
Obese 3334.1 22.1 0.4 11.4 0.8 0.9 2.3 2.4
Parity

No 3179.8 18.6 0.4 11.2 0.8 0.6 2.3 1.6
Yes 3346.7 22.8 0.5 12.1 0.9 1.0 2.0 2.6
Missing 3531.7 21.3 0.4 12.7 1.5 0.5 1.2 1.7
Maternal Income

<$15,000 3314.6 25.6 0.5 13.3 1.0 0.8 2.1 2.6
$15,000-$29,999 3335.8 19.6 0.5 10.5 0.9 0.8 2.3 1.6
$30,000+ 3346.8 19.9 0.5 9.6 0.9 1.2 2.0 2.9
Don't know 3231.2 20.8 0.4 13.2 0.8 0.7 1.9 2.1
Smoking History

Never 3308.3 21.2 0.5 11.8 0.9 0.7 2.0 2.3
Ever 3247.1 22.0 0.50 12.1 0.8 1.4 2.1 2.3
Notes: Spearman correlation coefficients
a
and Kruskal-Wallis one-way analysis of variance
b
; * = < 0.05; PM2.5 =
particulate matter with an aerodynamic diabetes less than 2.5µm; SHS = secondhand smoke; test results were without
missing or don't know level.

101
Table S.3.2. Spearman’s Correlation Coefficients for High-Loading Components of the Six Personal
PM 2.5 Sources (N = 201).

Table S.3.3. Effect Estimates of Major Personal PM 2.5 Sources on Birthweight in Full Term Births Only.
Main Model (N = 180) Outliers Included (N = 183)
Model β 95% CI β 95% CI
Traffic 6.7 -63.9 77.2 11.0 -62.2 84.2
Secondhand Smoke -3.6 -62.3 55.1 -11.8 -72.1 48.5
Aged Sea Salt -70.4 -146.9 6.2 -46.1 -119.8 27.7
Fresh Sea Salt -92.0 -196.2 12.2 21.8 -37.7 81.3
Fuel Oil 16.5
-41.2
74.1 28.0 -31.4 87.3
Crustal -16.2 -106.4
74.0
74.3 16.4 132.1
Notes: CI = confidence intervals; PM2.5 = particulate matter with an aerodynamic diabetes less than 2.5µm; β = change in
birthweight per 1 SD increase in pollutant.

Traffic
Secondhand
Smoke
Aged and Fresh
Sea Salt
(combined)
Fuel Oil Crustal

BC Zn Ba ETS BrC Br S Na Mg Cl V Cu Ni Al Ti Si Ca
BC 1.00

Zn 0.40 1.00

Ba 0.42 0.53 1.00

ETS -0.58 -0.17 -0.12 1.00

BrC -0.12 0.15 0.22 0.66 1.00

Br 0.28 0.25 0.29 0.05 0.23 1.00

S -0.08 0.05 0.07 0.09 0.06 0.25 1.00

Na -0.21 -0.08 -0.08 0.13 -0.02 0.09 0.64 1.00

Mg
-0.07 0.00 0.07 0.08 0.08 0.13 0.56 0.86 1.00

Cl -0.24 -0.01 -0.06 0.13 -0.03 -0.02 -0.07 0.46 0.43 1.00

V 0.01 -0.02 0.05 0.06 0.08 0.10 0.06 0.05 0.05 -0.01 1.00

Cu 0.26 0.24 0.39 -0.01 0.23 0.08 -0.11 -0.13 0.02 0.03 0.15 1.00

Ni 0.01 -0.06 0.14 0.11 0.15 0.05 -0.02 0.08 0.14 0.10 0.10 0.64 1.00

Al 0.29 0.34 0.34 -0.09 0.14 0.34 0.07 -0.01 0.18 0.04 0.04 0.26 0.12 1.00

Ti 0.34 0.47 0.48 -0.06 0.21 0.26 0.00 -0.08 0.13 0.03 -0.10 0.36 0.12 0.47 1.00

Si 0.37 0.56 0.49 -0.06 0.26 0.30 0.04 -0.03 0.21 0.09 0.04 0.44 0.23 0.58 0.65 1.00

Ca 0.35 0.49 0.53 -0.09 0.20 0.32 0.09 0.09 0.33 0.16 0.04 0.37 0.17 0.62 0.66 0.76 1.00
Notes. PM2.5 = particulate matter with an aerodynamic diabetes less than 2.5µm; bolded = p-value < 0.05; ETS = environmental tobacco smoke; BrC = brown
carbon; BC = black carbon.

102
CHAPTER 4
A COMPARISON OF MEASURED AIRBORNE AND SELF-REPORTED SECONDHAND
SMOKE EXPOSURE IN THE MADRES PREGNANCY COHORT STUDY
ABSTRACT
Background: Secondhand smoke (SHS) exposure during pregnancy is associated with several adverse
birth outcomes, including reduced birthweight and preterm birth. Questionnaires are commonly used to
assess SHS exposure; however, the wording of questions and their ability to capture true exposure can vary,
limiting researchers’ ability to harmonize SHS measures. Therefore, we compared the association of SHS
self-reported exposure in the MADRES pregnancy cohort with measurements of environmental tobacco
smoke (ETS) in particulate matter (PM 2.5) personal samples collected in the 3
rd
trimester.
Methods: We measured SHS on 48-hour integrated personal PM 2.5 Teflon filters collected from 204
pregnant women using an optical absorption approach. Self-reported SHS related to presence, intensity,
and duration of exposure was ascertained via a 3
rd
trimester questionnaire and an exit survey at the time of
personal monitoring. Descriptive statistics were calculated for ETS measures overall and by key
demographics and environmental factors. Analysis of variance tests were conducted to test group
differences in ETS concentrations by self-reported SHS exposure.
Results: Participants were predominately Hispanic (81%), with a mean (SD) age of 28.2 (6.0) years.
Geometric mean (SD) personal ETS concentrations were 0.14 (9.41) µg/m
3
. There was a significant
difference in mean ETS by education (p = 0.015), with participants with lower education having higher
measured ETS. Mean ETS concentrations differed by reported time with windows open (none/a little of the
time: 0.10 µg/m
3
vs. most/all of the time: 0.19 µg/m
3
; p = 0.047). There was no association between
measured ETS and self-reported SHS exposure; however, asking about the number of smokers nearby in
the 48-hour monitoring period was most correlated with measured ETS (Two+ smokers: 0.30 µg/m
3
vs.
One: 0.12 µg/m
3
and Zero: 0.15 µg/m
3
; p = 0.230).
Conclusion: Overall, self-reported SHS exposure was not associated with measured airborne ETS in
personal PM 2.5 samples.

103
INTRODUCTION
Secondhand smoke (SHS) exposure results in over 600,000 yearly deaths globally, including from
cancer (predominantly lung cancer), heart disease, and asthma (Oberg et al., 2011). In 2006, the United
States (U.S.) Surgeon General declared that there is no safe level of SHS (U.S. Department of Health and
Human Services, 2006), with widespread exposure and numerous adverse health risks, including
developmental effects and respiratory issues in infants and children (OEHHA, 2006). In pregnancy, SHS
exposure is associated with reduced birthweight (Hawsawi et al., 2015; Pogodina et al., 2009), preterm
births (Hoyt et al., 2018), congenital malformations (Zheng et al., 2019), and infant mortality (United States
Surgeon General, 2014). Possible biological mechanisms for the effects of SHS in pregnancy include
oxidative stress (Argalasova et al., 2019), metabolic and endocrine disruption (Flouris et al., 2010), and
increased secretion of the inflammatory cytokines IL-1β and TNF- α, leading to lower placental weight
(Niu et al., 2016).
SHS is a mixture of more than 7,000 chemical components released from both sidestream and
mainstream smoke from the burning of tobacco products including cigarettes, cigars, and pipes (US EPA,
2014). In addition to SHS that persists in the environment, thirdhand smoke (THS) is residual contamination
from tobacco smoke that lingers in rooms on surfaces or furniture, clothes, and hair after smoking and can
become re-volatized or resuspended (Jacob et al., 2017). Environmental tobacco smoke (ETS) is a term that
is often used as a synonym of SHS; however, it is also used to describe the combination of secondhand and
thirdhand smoke (THS) (Dede and Cinar, 2016). SHS concentration indoors is impacted by the number of
smokers, the pattern of smoking, and other factors including ventilation and infiltration (Apte et al., 2004;
Fabian et al., 2016; Jaakkola and Jaakkola, 1997).
Worldwide, as estimated 33-35% of nonsmokers are exposed to SHS, with numbers in children
estimated to be 40-50% (Oberg et al., 2011; Pirkle et al., 1996), although individual studies have reported
much higher prevalence of exposure (Fuentes-Leonarte et al., 2015; Wang et al., 2009). Actual exposures
in general may be much higher than reported, as self-reported exposure tends to be lower than more
objective biomarker and monitor-based exposure assessment. For example, researchers using data from the

104
National Health and Nutrition Examination Survey (NHANES) found that self-reported exposure rates in
children and adolescents were 14.6-23.7% and 15.2-23.4%, compared to urinary cotinine measured rates
of 50.9-65.2% and 47.0-63.2%, respectively (Max et al., 2009). Such discrepancies have also been seen in
pregnancy studies (Argalasova et al., 2019; Ashford et al., 2010), where while 72% of participants
responded no exposure on a self-reported questionnaire, almost all participants had detectable levels of
cotinine, indicating SHS exposure (DeLorenze et al., 2002). Therefore, questions remain around the actual
prevalence of SHS exposure in various populations and contexts and in the ability of various measures or
instruments at accurately capturing or estimating exposures in health studies.
The most common SHS exposure assessment approach is via questionnaire, which predominately
asks participants about the presence of smokers in their vicinity, and to a lesser extent, the intensity and
duration of smoke exposure (Avila-Tang et al., 2013). However, researchers found that the heterogeneity
of these questions was high and just 17.9% (~35% in pregnancy studies) of study questions were validated
(Pérez-Ríos et al., 2013). While studies generally support the validity of self-reported assessment (Kaufman
et al., 2002; Kraev et al., 2009; Vartiainen et al., 2002; Wipfli et al., 2008), there is evidence of exposure
misclassification (Cummings et al., 1990; Mourino et al., 2021; Woodruff et al., 2003). Questionnaires may
also be less accurate in pregnancy studies due social stigma, lack of exposure knowledge about exposure,
and recall bias may explain these discrepancies (Argalasova et al., 2019; DeLorenze et al., 2002; O’Connor
et al., 1995; Xiao et al., 2018). Additionally, the variation in question type, wording, and exposure windows,
makes it difficult to harmonize questionnaire data across studies, and validate questionnaire measures.
Exposure to SHS can be measured quantitatively using several approaches, ranging from stationary
measurements of airborne SHS-related chemicals concentrations (Leaderer and Hammond, 1991), to
personal monitoring (Lawless, 2004) and biomonitoring (Ashford et al., 2010; Chen et al., 2021; DeLorenze
et al., 2002). Measurements of tobacco smoke related chemicals in frequently occupied spaces or its
biomarkers are less likely to suffer from certain types of bias, including recall and social-desirability bias,
which may more readily occur with self-reported questionnaire data (Benowitz, 1996; Hsieh et al., 2011;
Jaakkola and Jaakkola, 1997). Biomarkers of exposure measure nicotine (and metabolites, including

105
cotinine) in blood, urine, saliva, and hair, and are considered the most accurate (gold standard) approach to
assess true exposure to nicotine since they are measuring an internal dose in the body, following intake and
closer to the biological effective dose. However, some biomarkers have short half-lives, such as cotinine
which ranges from 10-27 hours in body fluids (Benowitz, 1996). Nicotine in hair samples has a longer half-
life (Kim et al., 2014); however, they, like all biomarkers, cannot be used to trace where and when exposure
took place (Jaakkola and Jaakkola, 1997). Biomarkers are also invasive, requiring bio-specimens, such as
blood or urine. Additionally, researchers found that the cotinine per cigarette ratio during pregnancy was
significantly lower than in the postnatal period, suggesting that cotinine levels may be processed differently
during pregnancy compared to the post-natal period after accounting for smoking behaviors (Rebagliato et
al., 1998).
Air sampling can also be used to measure concentrations of smoke-related particle or vapor phase
chemicals in air. This can take the form of stationary monitoring in microenvironments or commonly
occupied indoor spaces (Kaur and Prasad, 2011; Semple et al., 2015; Williams et al., 1993) and personal
monitoring (Brook et al., 2011; Jenkins et al., 2001; Sloan et al., 2017). Additionally, nicotine badges can
be deployed as both stationary or personal samplers depending whether they are located in one location or
on the person, respectively (Eisner et al., 2001). Stationary monitors can be used to measure the
concentration of a marker of SHS using either passive (relying on gas diffusion) or active (using pumps to
draw air) sampling. Stationary monitoring can be a useful estimate of indoor air pollution, considering
people spend ~90% of their time inside (Habre et al., 2014). However, stationary monitors may fail to
account for individual-level exposure and do not sample the breathing zone, which is best captured by
personal monitoring (Gray et al., 2010). Personal air pollution monitoring is a highly effective way of
measuring individual-level exposures, especially if bio-specimens are not available, with prior personal
monitoring studies assessing SHS exposures by predominantly measuring nicotine (Eisner et al., 2001;
Jenkins and Counts, 1999; Muramatsu et al., 1984) and particulate matter (PM) (Thornburg et al., 2021;
Travers and Vogl, 2015; Zhang et al., 2020). However, personal monitoring of markers of SHS particles
too have limitations. For example, they are often costly and cumbersome to conduct, and they are often

106
short in duration (most commonly 24hr – 1 week). Additionally, while PM sampling may capture the
particle-phase of SHS, it does not capture other chemicals in the mixture, such as volatile or gas phase
components.
Common pollutants measured as markers of SHS concentration in air include particular matter
(Ratschen et al., 2018), nicotine (Martínez-Sánchez et al., 2018), various other metals and volatile organic
carbons (Apelberg et al., 2013), and carbon fractions of PM 2.5 (Yan et al., 2011). There are several
approaches to quantifying particulate markers of SHS (Ahmed et al., 2009; Apelberg et al., 2013; Benowitz,
1996; Yan et al., 2011). Optical methods are commonly used to quantify carbon fractions of PM 2.5 due to
their non-destructive and more affordable nature compared to thermal optical methods, and they have been
successfully used previously to estimate concentrations of black carbon (BC) and brown carbon (BrC) on
filters (Ahmed et al., 2009; Forder, 2014). Specifically, the multiwavelength approach used in this work
has been shown to have sufficient sensitivity to measure ETS (as a marker of SHS), BrC and BC from
personal PM 2.5, and has been successfully validated against other optical and thermal-optical approaches
(Yan et al., 2011).
Finally, in the prenatal air pollution exposure and health literature, there has been discussion around
whether to adjust for maternal smoking and SHS exposure during pregnancy as a covariate in health risk
models (Darrow, 2006; Jedrychowski et al., 2009; B. Ritz et al., 2007). The rationale for doing so seems to
rest on tobacco smoke exposure being a major risk factor for a given adverse health outcome or that it may
confound the relationship between the air pollutant and the outcome. However, most air pollution research
uses ambient or outdoor estimates of exposure rather than personal measures, which may be less susceptible
to biases from confounding by personal factors (Weisskopf and Webster, 2017). Therefore, adjusting for
SHS in health analyses may in fact depend on whether the exposure of interest is at the individual level,
obtained through personal measurement, or at residential or neighborhood level.
Therefore, the aims of this study were to first characterize personal PM 2.5-bound ETS
concentrations as surrogates of SHS exposure in the third trimester of pregnancy in the personal monitoring
sub-study of the Maternal and Developmental Risks from Environmental and Social Stressors (MADRES)

107
Pregnancy Cohort. Next, we evaluated the correlation between the optical ETS measurements and self-
reported SHS exposure from various questionnaires to determine which types or formulations of questions
best captured the variation in measured, personal PM 2.5-bound ETS. Finally, to determine whether SHS is
related to ambient or outdoor PM 2.5, and therefore shed light on discussions around whether it is necessary
to adjust for SHS as a potential confounder in outdoor air pollution and health studies, we investigated the
correlation of personal, PM 2.5-bound ETS with Personal PM 2.5 mass and Outdoor PM 2.5 mass
concentrations.
METHODS AND MATERIALS
Study Population
MADRES is an ongoing longitudinal cohort study of roughly 900 primarily Hispanic, low-income
pregnant women in Los Angeles County, CA, described previously (Bastain et al., 2019). Briefly, the aim
of MADRES is investigating the cumulative burden of environmental pollutants and behavioral,
psychosocial, and built environmental risk factors on maternal and infant health in a health disparities
population. Starting in November 2015 enrolled participants came from four prenatal care provider
partnerships with the following eligibility: 1) at least 18 years old, 2) fluency in either Spanish or English,
and 3) less than 30-weeks gestation at recruitment. Exclusion criteria for the study included: 1) a physical,
mental, or cognitive disability that would inhibit informed consent attainment, 2) HIV positive, 3) multiple
gestations, and 4) current incarceration.
The analysis for the present study encompasses data from a personal PM 2.5 exposure monitoring
sub-study within the MADRES cohort. Expecting mothers in the 3
rd
trimester of pregnancy were asked to
wear a personal monitoring device housed in a single-strap purse for 48-hours. This sub-study sample and
the larger MADRES cohort were comparable for demographic, birth outcomes, and outdoor air pollution
metrics. Additional exclusion criteria for the personal monitoring study component included smoking
households (where at least one currently active smoker lives permanently) to prevent PM 2.5 sampler
overloading. However, this criterion was not consistently applied across the study and was later removed.

108
Study procedures were approved by the USC Institutional Review Board (IRB) and all participants
completed written informed consent at first study visit.
Personal PM 2.5 and Environmental Tobacco Smoke (ETS) Exposure Monitoring
Personal PM 2.5 Exposure Sampling
Between October 2016 and February 2020, total personal PM 2.5 exposure was measured over an
integrated 48-hours in the 3
rd
trimester in a subset of 214 women from the larger MADRES cohort using a
custom sampling design. Participants were recruited during their 3
rd
trimester study visit to the University
of Southern California (USC) clinic. At this point, participants were provided the sampling apparatus
enclosed in a purse which consisted of a Gilian Plus Datalogging Pump (Sensidyne Inc.) and a Harvard
PM 2.5 personal environmental monitor (PEM) with a 37mm Pall Teflo filter affixed to the shoulder strap,
with the inlet located near the breathing zone. The pump was programmed to begin actively sampling (50%
cycle) at midnight the day after recruitment, with a flowrate of 1.8 liters per minute (LPM). Trained,
bilingual staff members provided instructions for correct usage, including a demonstration of how to wear
the device.
Instructions for proper use of the sampling device included: 1) wear the sampling device as much
as possible when going through daily activities, 2) make sure the inlet does not get covered by hair or
clothing and remains at the breathing zone, 3) protect device from water, heat, children, and pets.
Exceptions to the use requirement included: driving, showering, sleep, or otherwise unable to. During these
activities, participants were requested to place the device nearby, such as on a bedside table, or on the
passenger seat of a car, making sure to keep the device elevated from the ground and away from walls due
to concerns of sampling artifacts from resuspended duct or removal on surfaces, respectively. The sampling
device was programmed to shut off after the sampling period and was picked up the following day by study
staff where a brief exit survey was given to participants.
Sampling devices were taken to the USC Exposure Analytics lab to be processed by trained staff.
Filters from within the PEM were removed and placed within a dedicated chamber to equilibrate, then

109
weighed gravimetrically in temperature and relative humidity-controlled glove box using an MT-5
microbalance (Mettler Toledo, Inc.) to determine PM 2.5 mass concentrations. Filters were then stored to for
subsequent elemental (X-ray fluorescence analysis) and carbon fractions analysis (described below) at the
Research Triangle Institute International, Inc.
Environmental Tobacco Smoke (ETS) Concentrations
Filters were analyzed using a four-wavelength optical absorption technique at RTI International,
Inc., for concentrations of Black Carbon (BC), Brown Carbon (BrC) and Environmental Tobacco Smoke
(ETS). This method has been shown to perform well against other carbon measurement approaches (Yan
et al., 2011) and is described in more detail in Lawless et al. (2004). In this work, we used the ETS
measurement as a surrogate of SHS exposure. Briefly, this optical method leverages the absorption
properties across differing wavelengths of various carbon components. (Lawless, 2004; Yan et al., 2011).
Each carbon component has a different optical density at differing wavelengths, allowing for the calculation
of mass loadings of each component. For example, BC absorbs light across the spectrum, while ETS which
is characterized by a yellowish brown color, absorbs less infrared and much more ultraviolet light (Lawless,
2004). ETS mass concentrations are reported in μg/m
3
.
Questionnaire and Electronic Medical Record Variables
Data for this study was collected throughout follow-up from recruitment into MADRES until infant
birth, from a sequence of in-person and telephone interview questionnaires by trained staff in either English
or Spanish. This included a post-personal monitoring study exit survey (most time-aligned with the personal
exposure measurements), a 3
rd
trimester specific questionnaire, and data abstracted from electronic medical
records (EMR).
Maternal Demographic and Environmental/Home Characteristic Variables
Maternal demographics included the following: age (years), race/ethnicity (Hispanic, non-Hispanic
Black and non-Hispanic Other), education level (completed <12th grade, completed high school, some
college, completed college), household income (less than $15,000, $15,000 – 29,999, $30,000+, Missing),
marital status (employed, student, homemaker, unemployed), parity (parous, nulliparous, missing).

110
A priori environmental and household characteristics that may determine or influence ETS
concentrations were evaluated with the following variables: home type (single-family home (no joining
wall), 2-4 attached units, 5+ units, missing), window opening time (none and a little of the time vs. most
and all of the time), air conditioning use (AC; none of the time vs. a little, most, and all of the time), time-
spent indoors (none and a little of the time vs. most and all of the time), cooking smoke exposure (none of
the time vs. a little, most, and all of the time), candle or incense smoke exposure (none of the time vs. a
little, most, and all of the time). Several of the questionnaire variables were re-categorized, when necessary,
based on the distribution of the variable.
Self-Report Secondhand Smoke Exposure Questionnaire and EMR Variables
Self-reported smoking and SHS exposure questionnaire variables were accumulated from all
possible MADRES questionnaires to evaluate how different questionnaire type/wording and time-point
correlated with measured ETS concentrations. The personal monitoring exit survey consisted of the
following SHS questions: 1) Approximately how much of the time were you close to cigarette, cigar,
hookah or pipe smoke from people smoking nearby? 2) If greater than none, how many people were
smoking nearby? Within the 3
rd
trimester questionnaire, the following questions were asked: 1) Since we
last saw you/spoke to you in your first/second trimester, excluding e-cigarettes, has anyone else living in
your home smoked cigarettes, cigars or pipes inside the house? 2) Since we last saw you/spoke to you in
your first/second trimester, how many hours per day have you been exposed to cigarette, cigar or pipe
smoke because of smoking by others? Missing self-reported SHS exposure data in the 3
rd
trimester was
replaced with data from the 2
nd
followed by the 1
st
trimester questionnaires for <5% of participants. Finally,
participants’ EMRs provided the following SHS variable: 1) History of any second-hand smoking
exposure? Only self-reported SHS exposure was evaluated in this study since virtually all expecting
mothers reported being non-smokers. In keeping with prior studies (Avila-Tang et al., 2013; Pérez-Ríos et
al., 2013), SHS exposure variables were grouped into three general domains: 1) presence of a smoker

111
nearby, 2) intensity of exposure (number of smokers nearby) and 3) duration of exposure (time exposed to
smoking). Several of the self-reported SHS variables underwent re-categorization based on the distribution.
Outdoor Residential PM 2.5
Daily outdoor residential PM 2.5 estimates were assigned to participants’ locations using an inverse-
distance weighted spatial interpolation approach from ambient air quality monitoring data (United States
Environmental Protection Agency Air Quality System), with a time-weighting method used for subjects
with multiple residences.
Statistical Analysis
PM 2.5 and ETS by Key Characteristics
Personal PM 2.5 and ETS concentrations (μg/m
3
) were log-transformed for this analysis due to right
skew, with a small offset added to zero values for transformation purposes. Geometric means and standard
deviations were calculated for personal PM 2.5 and ETS concentrations in μg/m
3
. Next, personal PM 2.5 and
ETS concentrations were summarized by sample demographic characteristics and environmental factors
and analysis of variance (ANOVA) tests were used to determine whether exposures varied by these factors.
Log-transforming ETS allowed for the use of parametric tests, which have greater statistical power on
skewed data than non-parametric tests, such as Kruskal-Wallis (Rasmussen and Dunlap, 1991).
PM 2.5 and ETS by Self-Reported SHS Exposure
Measured airborne personal PM 2.5 and ETS concentrations were then compared to self-report SHS
exposure to investigate whether questions related to smoker proximity, smoking intensity, and duration of
ETS exposure were useful for capturing the variation in ETS measurements. Similarly, ANOVA tests were
used to test whether measured ETS exposure differed by levels of self-report SHS exposure for all these
questions. No statistical test was conducted on the EMR self-reported ETS question due to small cell counts
(see discussion).
Additional ETS metrics were also considered in similar analyses to determine whether results are
sensitive to the definition of ETS exposure. The ratio of ETS to personal PM 2.5 concentration was calculated

112
to account or adjust for variable PM 2.5 mass concentrations. Additionally, a binary variable was created
based on the 75
th
percentile ETS cut-point (0.5 μg/m
3
), which was similar to a prior study that used 0.4
μg/m
3
as a cut-point for ETS exposure (Brook et al., 2011). However, results are not included as both these
analyses did not reveal any additional information about the relationship between measure ETS and self-
reported SHS.
Personal ETS vs. Personal and Outdoor PM 2.5 Exposure
To provide insight into the rationale for adjusting for SHS in air pollution health analyses,
Spearman correlations of personal ETS concentrations with personal PM 2.5 mass concentrations and
outdoor, residential PM 2.5 concentrations (in the same 48 hours as the monitoring period) were calculated
given the distribution of ETS was right-skewed.
Statistical significance was determined with an alpha level of 0.05 for all tests. Models were
assessed for violations of model diagnostics and influential points. Analyses were conducted using SAS
v9.4 and JMP 16 Pro (SAS Institute, Inc., Cary, NC, USA).
RESULTS
Descriptive Statistics
From the original 214 participants in the MADRES personal monitoring sub-study, eight
participants were removed due to erroneous or incomplete personal PM 2.5 measurements. An additional two
women were removed due to missing ETS measurement, resulting in a 204 final sample that was included
in the analysis. This sample had a mean (SD) age of 28.16 (6.01) years, was predominately Hispanic
(80.69%), with around 63% having had a prior pregnancy. Additionally, over 55% of participants had
completed high school or less, and over 40% of the women reported a household income less than $30,000
a year. The geometric mean and standard deviation for personal PM 2.5 and ETS concentrations were 17.70
(1.88) and 0.14 (9.41) μg/m
3
, respectively.
PM 2.5 and ETS by Key Characteristics
Differences in personal PM 2.5 and ETS concentrations were observed by demographic
characteristics as presented in Table 4.1. Education level was statistically significantly associated with

113
measured airborne ETS concentrations, with lower educated participants having greater ETS concentrations
(<12
th
grade: 0.16 μg/m
3
, completed high school: 0.27 μg/m
3
, some college: 0.08 μg/m
3
, completed college:
0.10 μg/m
3
; p = 0.015). There was also marginally significant difference in personal PM 2.5 by education
level (p = 0.093), with the lowest level (less than 12
th
grade) experiencing 19.42 μg/m
3
vs 13.68 μg/m
3
for
the highest education level (completed college). While race/ethnicity did not meet statistical significance
(p = 0.162), non-Hispanic Black participants were exposed to higher geometric mean ETS of 0.28 μg/m
3
,
compared to 0.14 μg/m
3
and 0.07 μg/m
3
for Hispanic and non-Hispanic Other participants, respectively.
Table 4.1. Summary of Study Participants by Measured Personal PM 2.5 and ETS Concentrations.
PM 2.5 ETS
Characteristic
Mean (SD)
or n (%)
Geometric
Mean (SD)

(µg/m
3
)
P-value
a

Geometric
Mean (SD)

(µg/m
3
)
P-value
a

Full Sample 204 17.70 (1.88) 0.14 (9.41)
Maternal Age (years) 28.16 (6.01) - -
Race 0.244 0.162
Hispanic 163 (80.69%) 18.03 (1.86) 0.14 (8.46)
Black, non-Hispanic 22 (10.89%) 18.66 (1.95) 0.28 (13.25)
Other, non-Hispanic 17 (8.42%) 13.85 (1.98) 0.07 (15.89)
Education 0.093 0.015
<12th grade 49 (24.26%) 19.42 (1.92) 0.16 (6.24)
Completed High School 64 (31.68%) 18.28 (1.77) 0.27 (8.79)
Some College 58 (28.71%) 18.12 (2.04) 0.08 (12.05)
Completed College 31 (15.35%) 13.69 (1.68) 0.10 (9.77)
Maternal Income 0.105 0.615
Less than $15,000 41 (20.30%) 20.73 (1.80) 0.18 (15.52)
$15,000-$29,999 46 (22.77%) 16.98 (1.87) 0.11 (8.95)
$30,000+ 41 (20.30%) 15.51 (1.90) 0.14 (6.99)
Don't know 74 (36.63%) 17.91 (1.92) 0.15 (8.82)
Parity 0.117 0.749
Yes 128 (62.75%) 18.62 (1.86) 0.15 (8.87)
No 68 (33.33%) 16.08 (1.92) 0.14 (11.00)
Missing 8 (3.92%) 17.93 (1.90) 0.12 (7.63)
Marital Status 0.731 0.654
Married 55 (26.96%) 18.53 (2.03) 0.13 (8.47)
Living together 88 (43.14%) 17.81 (1.65) 0.17 (10.02)
Single 47 (23.04%) 16.77 (2.09) 0.13 (8.47)

114
Missing 14 (6.86%) 17.09 (2.07) 0.09 (14.75)
Employment Status 0.812 0.662
Employed 86 (42.36%) 16.99 (1.88) 0.12 (9.23)
Student 21 (10.34%) 17.35 (1.89) 0.15 (14.14)
Homemaker 55 (27.09%) 18.82 (1.94) 0.20 (7.60)
Unemployed 41 (20.20%) 18.19 (1.84) 0.13 (10.04)
Notes: PM 2.5 = particulate matter with an aerodynamic diameter less than 2.5µm; ETS = environmental
tobacco smoke; SD = standard deviation;
a
analysis of variance test; bolded = statistically significant at
p-value < 0.05; p-value are for tests without missing or don't know levels.
There were observable differences in both personal PM 2.5 and ETS concentrations by environmental
factors including housing characteristics (Table 4.2.). Personal PM 2.5 concentrations were not statistically
different by window opening behavior (as reported in the exit survey); however, ETS concentrations were
significantly different (p = 0.047) with higher concentrations in participants that reported opening their
windows most and all of the time vs. none and a little of the time. Marginally significant differences in both
personal PM 2.5 (p = 0.099) and ETS (p = 0.059) concentrations were found by home type (missing level
removed for statistical test), with single-family homes having the lowest concentrations compared to
multiunit residences; however, there was no clear trend with increasing number of residential units. Finally,
we found a statistically significant association between self-reported candle or incense smoke exposure and
personal PM 2.5 mass concentrations (a little, most, and all of the time 20.62 μg/m
3
vs. 16.67 μg/m
3
none of
the time; p = 0.036) but no differences in ETS concentrations.
Table 4.2. Summary of Environment/Household Characteristics by Measured Personal PM 2.5 and ETS
Concentrations.
PM 2.5 ETS
Question n (%)
Geometric
Mean (SD)

(µg/m
3
)
P-
value
a

Geometric
Mean (SD)

(µg/m
3
)
P-
value
a

Home Type 0.099 0.059
Which best describes the home in which you currently live most of the time?
Single-family home 71 (34.80%) 15.40 (1.74) 0.10 (9.21)
2-4 units 48 (23.53%) 19.37 (1.86) 0.26 (5.74)
5+ units 67 (32.84%) 18.40 (2.01) 0.13 (12.29)
Missing 18 (8.82%) 20.90 (1.90) 0.16 (9.66)
Window Opening 0.564 0.047

115
How much of the time were windows (or porch/balcony doors if applicable) open in your home,
when you were there with the sampler?
Most and all of the
time 125 (61.58%) 17.22 (1.75) 0.19 (8.24)
None and a little of
time 78 (38.42%) 18.14 (2.05) 0.10 (10.45)
Air Conditioner Use 0.930 0.642
How much of the time was the air conditioner used in your home, when you were there with the
sampler?
A little, most, or all
of the time 54 (26.60%) 17.45 (1.87) 0.13 (7.75)
None of the time 149 (73.40%) 17.61 (1.87) 0.15 (9.81)
Cooking Smoke 0.750 0.255
How much of the time were you close to smoke or fumes from cooking (yourself, or nearby cooking
by someone else) e.g. burnt toast, barbeque, stir fry, etc.?

A little, most, or all
of the time 80 (39.41%) 17.26 (1.78) 0.12 (9.64)
None of the time 123 (60.59%) 17.77 (1.93) 0.17 (8.90)
Candle/Incense Smoke 0.036 0.235
How much of the time were you close to smoke from candles or incense burning nearby?
a

A little, most, or all of
the time 50 (24.63%) 20.62 (1.81) 0.11 (16.35)
None of the time 153 (75.37%) 16.67 (1.87) 0.16 (7.36)
Commute (in car, bus,
or train) 0.172 0.184
Did you commute in a car, bus or train on roadways?
Yes 174 (85.71%) 18.00 (1.84) 0.14 (9.90)
No 29 (14.29%) 15.16 (2.02) 0.24 (5.28)
Notes: PM 2.5 = particulate matter with an aerodynamic diameter less than 2.5µm; ETS =
environmental tobacco smoke; SD = standard deviation;
a
analysis of variance test; bolded =
statistically significant at p-value < 0.05; p-value are for tests without missing or don't know levels.
PM 2.5 and ETS by self-reported SHS exposure
Table 4.3. shows the relationship between personal PM 2.5 ETS concentrations and self-reported
SHS exposure questionnaire variables. None of the self-reported SHS questions was statistically
significantly associated with PM 2.5 or ETS concentrations; however, number of people smoking nearby
(referring to intensity of exposure) seemed to be the most correlated with measured airborne ETS
concentrations. Specifically, being exposed to two or more smokers in the 48-hour monitoring period was
associated with ETS concentrations (geometric mean (SD) 0.30 (9.72) μg/m
3
) compared one smoker (0.12
(8.95) μg/m
3
) and zero smokers (0.15 (9.26) μg/m
3
). Increasing duration of SHS exposure (question

116
referring to how much of the time were you close to smoke) did not trend with higher ETS concentrations
(Table 4.3.). History of SHS exposure determined from EMR abstraction was not tested statistically due to
low cell counts, but geometric mean ETS was not noticeably different between exposed, unexposed, and
not recorded.
Table 4.3. Summary of Measured Personal PM 2.5 and ETS Concentrations by Self-Reported SHS Exposure
Questionnaire Responses.

PM 2.5 ETS
Question n (%)
Geometric
Mean (SD)

(µg/m
3
)
P-value
a

Geometric
Mean (SD)

(µg/m
3
)
P-
value
a

Presence of Smoker Nearby or in Residence
0.445 0.942
Approximately how much of the time were you close to cigarette, cigar, hookah or pipe smoke from
people smoking nearby?
b

A little, most, or all of the
time 80 (39.60%) 16.89 (2.04) 0.15 (9.37)

None of the time 122 (60.40%) 18.09 (1.75) 0.15 (9.26)

0.435

0.870
Since we last saw you/spoke to you in your first/second trimester, excluding e-cigarettes, has anyone
else living in your home smoked cigarettes, cigars or pipes inside the house?
c

Yes 11 (5.42%) 20.53 (1.67) 0.16 (12.20)

No 192 (94.58%) 17.60 (1.90) 0.14 (9.38)

-
-
History of any second-hand smoking exposure?
d

Yes 3 (1.47%) 15.69 (1.56) 0.08 (55.93)
No 184 (90.20%) 17.66 (1.90) 0.15 (9.34)
Not Recorded 17 (8.33%) 18.59 (1.84) 0.10 (8.09)
Intensity of SHS Exposure

0.432

0.230
If greater than none, how many people were smoking nearby?
b

2+ 27 (13.37%) 17.00 (1.77)

0.30 (9.72)

1 40 (19.80%) 15.69 (2.14) 0.12 (8.95)

0 122 (60.40%) 18.09 (1.75) 0.15 (9.26)

Don't Know 13 (6.44%) 20.88 (2.31) 0.07 (7.99)

Duration of SHS Exposure

0.248

0.831
Since we last saw you/spoke to you in your first/second trimester, how many hours per day have you
been exposed to cigarette, cigar or pipe smoke because of smoking by others?
c

2+ hours 15 (7.43%) 21.23 (1.70) 0.15 (7.27)
1-2 hours 17 (8.42%) 14.61 (2.27) 0.20 (14.21)

117
0-1 hours 170 (84.16%) 17.81 (1.86) 0.14 (9.40)

Notes: PM 2.5 = particulate matter with an aerodynamic diameter less than 2.5µm; ETS =
environmental tobacco smoke; SD = standard deviation;
a
analysis of variance test; bolded =
statistically significant at p-value < 0.05; p-value are for tests without missing or don't know levels;
b
=
exit survey;
c
= 3
rd
trimester questionnaire;
d
= abstracted from electronic medical records.
Personal ETS vs. Personal and Outdoor PM 2.5 exposure
Table 4.4. shows a marginally statistically significant and weak correlation between personal ETS
and PM 2.5 concentrations (Spearman ρ: 0.13; p = 0.061); however, there was no correlation between
measured, personal ETS and ambient, residential PM 2.5 concentration (ρ: -0.04; p = 0.609). Personal and
ambient PM 2.5 for the same 48-hour sampling period were also not statistically significantly correlated (ρ:
0.09; p = 0.215).
Table 4.4. Bivariate Analysis of ETS by Personal and Ambient PM 2.5.
Pollutant
Spearman ρ (p-
value)
ETS Personal PM 2.5 Ambient PM 2.5
ETS 1.00
Personal PM 2.5 0.13 (0.061) 1.00
Ambient PM 2.5 -0.04 (0.609) 0.09 (0.215) 1.00
Notes: PM 2.5 = particulate matter with aerodynamic diameter less than 2.5µm; ETS = environmental
tobacco smoke; Average ambient PM 2.5 corresponds with the same 48-hour personal exposure
sampling period.
DISCUSSION
In this analysis, we quantified and evaluated an objective measure of airborne ETS in the 3
rd

trimester, ascertained through a multiwavelength optical absorption technique, to better understand SHS
exposure levels and how they differed across key characteristics (demographic and environment/household
factors) in a health disparities population in Los Angeles, CA. Next, we conducted an analysis of how well
five SHS questions from different time-points across the 3
rd
trimester and wording type explained variation
in personal ETS exposure. To the best of our knowledge, this is the first time this optically derived ETS
measure from a personal PM 2.5 exposure monitoring study has been compared to self-reported SHS
questions during pregnancy. Additionally, this is an important question because the effects of SHS are still

118
prominent in society, and exposure misclassification is a major concern in health analyses, particularly in
pregnancy studies.
Overall, self-reported SHS exposure questions in various timepoints across the 3
rd
trimester of
pregnancy were not associated with measured airborne ETS in personal PM 2.5 samples. However, the
question regarding the number of smokers nearby was most correlated with ETS measurements compared
to questions that asked about the presence of a smoker nearby or duration of time exposed to smoke. A
prior study also found that number of smokers (in a household) was an important factor to account for
variation in SHS exposure using serum cotinine levels (Kaufman et al., 2002). Additionally, in this study,
out of 204 participants, airborne ETS concentrations was present in 186 (91%) participants. Together, this
highlights that in the current absence of a standardized battery of SHS exposure questionnaire measures for
pregnant women, there is a need to validate self-reported SHS questionnaires for use in health analyses to
minimize the risk of obtaining biased results. While many studies have found concordance between self-
reported and objective SHS exposure measures in pregnancy studies (Christensen et al., 2004), others have
found a great amount of discordance just as we did (Argalasova et al., 2019; DeLorenze et al., 2002;
O’Connor et al., 1995; Xiao et al., 2018). Pérez-Ríos et al. (2013) found that just 17.9% of epidemiological
studies (~35% in pregnancy studies) validated their SHS exposure assessment, potentially introducing
exposure misclassification and potential bias in these analyses. SHS exposure has been shown to be usually
underreported or underestimated in validation studies, which might potentially result in noisier or less
precise health risk estimates attenuated health effects (Carroll, 2005).
We also did not see any indication of the ETS measure capturing or correlating with other
combustion sources in the home such as cooking or candle and incense burning, which adds to the
confidence in our analysis and assumption that ETS is correlating with SHS exposure. However, it is
possible that this measure is more sensitive to more very fresh or more proximal or nearby SHS exposure
given the optical method was developed or optimized in chamber studies burning cigarettes.
Furthermore, there were noticeable differences in responses from participants for self-reported SHS
exposure for similar question type, taken from different time point. For example, 80 participants answered

119
that they were close to people smoking nearby at least some of the time during the 48-hour personal
sampling period, while 11 participants stated they lived in a home with a smoker that smoked inside, and
just 3 were recorded as being exposed to SHS in their EMR. This highlights that question type, wording of
question, and time window may all impact the degree of exposure misclassification when self-report is
used, and SHS exposure ascertained through EMR abstraction may not be sufficient or well representative.
Exposure to SHS is not always overt and participants may simply be unaware of their exposure,
likely resulting in non-differential misclassification; however, the social stigma associated with smoking
(particularly during pregnancy) may make smokers and those around smokers underreport, weakening
health effect estimates. Woodruff et al. (2003) reported modest associations in Latinos between parental
SHS report and both nicotine and cotinine hair samples of their child, suggesting there may be
racial/ethnicity differences in the sensitivity of self-report questions. However, we were unable to stratify
this analysis by race/ethnicity due to low sample sizes of non-Hispanic Black, and non-Hispanic Other.
Currently there is no agreed upon cut point for measured airborne ETS to indicate SHS exposure
as with cotinine in biospecimens (Benowitz et al., 2009). Prior studies that used a similar approach
measured ETS concentrations from “negligible” (no concentration given) (Sloan et al., 2017) to 4.0 (12.7)
µg/m
3
(Brook et al., 2011), with ETS concentrations exhibiting a right skew to the distribution. Brook et al.
(2011) study was in both male and female adults within Wayne County, MI, with a larger percent of African
Americans who were also exposed to greater concentrations of ETS similar to our study (despite our smaller
sample size). Additionally, our study is a pregnancy study where participants may be more likely to avoid
smoking and/or being around other smokers. There were also methodological differences especially related
to sample integration time and wear mode between our study and the two studies mentioned here. Compared
to our 48-hour integrated personal exposure monitor located near the shoulder of participants, Brook et al.
(2011) used a monitoring vest over 24 hours, while Sloan et al. (2017) attached a personal exposure monitor
to a stroller or diaper bag over a 7-day period. It is unclear whether these differences might explain
differences with the results in this present study.

120
We also found that individuals with lower educational attainment had significantly higher ETS
exposures, which is similar to other studies (Argalasova et al., 2019; Levine et al., 2013). Lower education
of parents was also associated with higher SHS exposure in their children (Bolte et al., 2008; Protano et al.,
2019). Interestingly, in this analysis participants that had completed high school that had the highest
measured airborne ETS exposure, compared to those with some college or greater, and even those who
didn’t complete high school. Evaluating whether participants were U.S. vs. non-U.S. born did not help
explain this finding. Additionally, while no significant differences were observed by race/ethnicity, non-
Hispanic Blacks had geometric means twice and four times higher than Hispanic and non-Hispanic Others,
respectively. This is in keeping with prior studies that found that African-Americans are exposed to higher
levels of SHS across all ages (Shastri et al., 2021; Tsai, 2018; Wilson et al., 2005). This is similar to others
who found non-Hispanic blacks had the highest prevalence of SHS (Tsai, 2018). This may be due to
increased targeted marketing of tobacco products at African-Americans (Primack et al., 2007).
Smoking ban policies in the workplace can be effective at reducing exposure to SHS (Wynne and
Bonevski, 2018). We did not find any significant differences between employed and unemployed
individuals, which is noteworthy considering the workplace is still often described as a place where SHS
exposure is prevalent in the literature (Max et al., 2009; Park et al., 2019), potentially showing the benefits
of smoke free public spaces. Even though significant differences were not observed, homemakers had
higher measured airborne ETS compared to employed, students, and unemployed participants, although it
is not clear why.
Home characteristics are thought to play an important role in exposure to SHS (Apte et al., 2004;
Price et al., 2006), and we observed that both home type and window opening played a role in SHS exposure
for pregnant women in our study. Individuals who lived in homes where window opening was high had
statistically significantly higher measured airborne ETS exposure, highlighting how ventilation may impact
exposure by potentially providing an avenue for SHS to enter the home potentially from nearby units or
residents. Additionally, the type of home participants lived in was marginally associated with airborne ETS,
with the multi-unit residents experiencing higher ETS compared to those who resided in single-family

121
homes, which has been observed by others (Bonevski et al., 2014). This is possibly due to leakage and
infiltration from neighboring units, which is increased in buildings and units in buildings with multiple
units (Price et al., 2006).
Finally, researchers have found that maternal active smoking and passive smoking (synonym for
SHS exposure) do not appear to be potential confounders in ambient air pollution and adverse birth
outcomes analyses (Darrow, 2006; Beate Ritz et al., 2007). In this analysis we provided further supporting
evidence that personal exposure to ETS is not associated with ambient PM 2.5 concentrations, highlighting
that SHS exposure may not meet the criteria for traditional confounding. Decisions to include SHS exposure
as a possible covariate in health analyses should take this into account. There was a weak significant
correlation between personal ETS exposure and personal PM 2.5 mass concentration, suggesting that
adjustment for SHS in personal PM 2.5 studies may be more necessary considering personal behaviors are
more likely to be associated with personal rather than ambient air pollution (Weisskopf and Webster, 2017).
Therefore, adjustment for SHS exposure as a potential confounder may depend on the research question
and more specifically, whether total personal PM 2.5 or ambient (outdoor origin) PM 2.5 is the exposure of
interest in the health analysis.
There are several strengths of this analysis. We were able to use personal exposure monitoring and
a multiwavelength optical absorbance technique to objectively describe SHS exposure in a health disparities
population. This analysis provides insights into the SHS exposures of pregnant women in Los Angeles, CA,
highlighting that even in a cohort of largely non-smoking households, SHS is ubiquitous. Additionally, the
MADRES cohort is a well characterized study with a host of information on demographic and environment
factors, making this an ideal setting to assess how well objective measurements correlated with self-reported
SHS exposure questionnaire measures from several time-points, different wording, and capturing different
aspects or domains of the exposure (e.g., presence of a smoker vs. number of smokers nearby, etc.).
The sample size of this study is a potential limitation, which while small for a population-based
study is quite large for personal exposure monitoring studies (Dadvand et al., 2012; Sarnat et al., 2000; Suh
and Zanobetti, 2010). Despite this limitation we were able to detect differences in measured airborne ETS

122
concentrations and education, home type, and window opening behavior. Next, the optical carbon analysis
method used is only capturing the particle phase of ETS in personal PM 2.5 and does not capture gas/volatile
phases of SHS or even thirdhand smoke or other more specific chemical tracers like nicotine. However,
this approach has performed well against other optical and optical-thermal approaches (Yan et al., 2011).
Finally, the sampling period was a single 48-hour sampling period in the 3
rd
trimester of pregnancy, which
may not reflect a typical exposure during pregnancy due to the short study period, or because behaviors
may change during the monitoring phase. We evaluated behavioral patterns, such as window use and time
activity patterns across study questionnaires, and found reasonable agreement between 3
rd
trimester and
exit survey at the end of the monitoring period, suggesting behaviors were generally similar across time
(results not shown).
Overall, this work highlights that SHS exposure is endemic in this group of largely lower income,
Hispanic pregnant women in Los Angeles, CA, even in households that avoid smoking. SHS is a known
risk factor for several adverse health outcomes; however, this study shows a poor correlation between self-
reported questions and measured “true” exposure, highlighting a pressing need for greater harmonization
of questionnaire-based SHS measures. Lastly, epidemiology studies should consider the research question
and exposure of interest when considering adjusting for SHS exposure as a potential confounder in air
pollution studies, as personal behaviors are more likely to be associated with personal rather than ambient
air pollution.

123
REFERENCES
Ahmed, T., Dutkiewicz, V.A., Shareef, A., Tuncel, G., Tuncel, S., Husain, L., 2009. Measurement of
black carbon (BC) by an optical method and a thermal-optical method: Intercomparison for four
sites. Atmospheric Environment 43, 6305–6311. https://doi.org/10.1016/j.atmosenv.2009.09.031
Apelberg, B.J., Hepp, L.M., Avila-Tang, E., Gundel, L., Hammond, S.K., Hovell, M.F., Hyland, A.,
Klepeis, N.E., Madsen, C.C., Navas-Acien, A., Repace, J., Samet, J.M., Breysse, P.N., 2013.
Environmental monitoring of secondhand smoke exposure. Tobacco Control 22, 147–155.
https://doi.org/10.1136/tobaccocontrol-2011-050301
Apte, M.G., Gundel, L.A., Dod, R.L., Russell, M.L., Singer, B.C., Sohn, M.D., Sullivan, D.P., Chang, G.-
M., Sextro, R.G., 2004. Indoor Measurements of Environmental Tobacco Smoke Final Report to
the Tobacco Related Disease Research Program (No. LBNL--49148, 823921).
https://doi.org/10.2172/823921
Argalasova, L., Zitnanova, I., Vondrova, D., Dvorakova, M., Laubertova, L., Jurkovicova, J., Stofko, J.,
Weitzman, M., Waczulikova, I., Simko, M., 2019. Self-Reported Exposure to ETS
(Environmental Tobacco Smoke), Urinary Cotinine, and Oxidative Stress Parameters in Pregnant
Women—The Pilot Study. Int J Environ Res Public Health 16, 1656.
https://doi.org/10.3390/ijerph16091656
Ashford, K.B., Hahn, E., Hall, L., Rayens, M.K., Noland, M., Collins, R., 2010. Measuring prenatal
secondhand smoke exposure in mother–baby couplets. Nicotine Tob Res 12, 127–135.
https://doi.org/10.1093/ntr/ntp185
Avila-Tang, E., Elf, J.L., Cummings, K.M., Fong, G.T., Hovell, M.F., Klein, J.D., McMillen, R.,
Winickoff, J.P., Samet, J.M., 2013. Assessing secondhand smoke exposure with reported
measures. Tob Control 22, 156–163. https://doi.org/10.1136/tobaccocontrol-2011-050296
Benowitz, N.L., 1996. Cotinine as a Biomarker of Environmental Tobacco Smoke Exposure.
Epidemiologic Reviews 18, 188–204. https://doi.org/10.1093/oxfordjournals.epirev.a017925
Benowitz, N.L., Schultz, K.E., Haller, C.A., Wu, A.H.B., Dains, K.M., Jacob, P., 2009. Prevalence of
Smoking Assessed Biochemically in an Urban Public Hospital: A Rationale for Routine Cotinine
Screening. Am J Epidemiol 170, 885–891. https://doi.org/10.1093/aje/kwp215
Bolte, G., Fromme, H., for the GME Study Group, 2008. Socioeconomic determinants of children’s
environmental tobacco smoke exposure and family’s home smoking policy. The European
Journal of Public Health 19, 52–58. https://doi.org/10.1093/eurpub/ckn114
Bonevski, B., Paul, C., Jones, A., Bisquera, A., Regan, T., 2014. Smoky homes: gender, socioeconomic
and housing disparities in second hand tobacco smoke (SHS) exposure in a large population-
based Australian cohort. Prev Med 60, 95–101. https://doi.org/10.1016/j.ypmed.2013.12.024
Brook, R.D., Bard, R.L., Burnett, R.T., Shin, H.H., Vette, A., Croghan, C., Phillips, M., Rodes, C.,
Thornburg, J., Williams, R., 2011. Differences in blood pressure and vascular responses
associated with ambient fine particulate matter exposures measured at the personal versus
community level. Occup Environ Med 68, 224–230. https://doi.org/10.1136/oem.2009.053991
Carroll, R.J., 2005. Measurement Error in Epidemiologic Studies, in: Encyclopedia of Biostatistics.
American Cancer Society. https://doi.org/10.1002/0470011815.b2a03082
Chen, M.M., Guo, S.-E., Yuan, C.-P., Okoli, C., Liao, Y.-C., 2021. Association between Self-Reported
Survey Measures and Biomarkers of Second-Hand Tobacco Smoke Exposure in Non-Smoking
Pregnant Women. International Journal of Environmental Research and Public Health 18, 9197.
https://doi.org/10.3390/ijerph18179197
Christensen, A.E., Tobiassen, M., Jensen, T.K., Wielandt, H., Bakketeig, L., Høst, A., 2004. Repeated
validation of parental self-reported smoking during pregnancy and infancy: a prospective cohort
study of infants at high risk for allergy development. Paediatr Perinat Epidemiol 18, 73–79.
https://doi.org/10.1111/j.1365-3016.2003.00520.x

124
Cummings, K.M., Markello, S.J., Mahoney, M., Bhargava, A.K., McElroy, P.D., Marshall, J.R., 1990.
Measurement of current exposure to environmental tobacco smoke. Arch Environ Health 45, 74–
79. https://doi.org/10.1080/00039896.1990.9935929
Dadvand, P., de, N.A., Triguero, -Mas Margarita, Schembari, A., Cirach, M., Amoly, E., Figueras, F.,
Basaga, ña X., Ostro, B., Nieuwenhuijsen, M., 2012. Surrounding Greenness and Exposure to
Air Pollution During Pregnancy: An Analysis of Personal Monitoring Data. Environmental
Health Perspectives 120, 1286–1290. https://doi.org/10.1289/ehp.1104609
Darrow, L.A., 2006. Maternal Smoking as a Confounder in Studies of Air Pollution and Infant Mortality.
Dede, C., Cinar, N., 2016. Environmental Tobacco Smoke and Children’s Health. Iran J Pediatr 26,
e5935. https://doi.org/10.5812/ijp.5935
DeLorenze, G.N., Kharrazi, M., Kaufman, F.L., Eskenazi, B., Bernert, J.T., 2002. Exposure to
environmental tobacco smoke in pregnant women: the association between self-report and serum
cotinine. Environ Res 90, 21–32. https://doi.org/10.1006/enrs.2001.4380
Eisner, M.D., Katz, P.P., Yelin, E.H., Hammond, S.K., Blanc, P.D., 2001. Measurement of environmental
tobacco smoke exposure among adults with asthma. Environ Health Perspect 109, 809–814.
https://doi.org/10.1289/ehp.01109809
Fabian, M.P., Lee, S.K., Underhill, L.J., Vermeer, K., Adamkiewicz, G., Levy, J.I., 2016. Modeling
Environmental Tobacco Smoke (ETS) Infiltration in Low-Income Multifamily Housing before
and after Building Energy Retrofits. Int J Environ Res Public Health 13, 327.
https://doi.org/10.3390/ijerph13030327
Flouris, A.D., Vardavas, C.I., Metsios, G.S., Tsatsakis, A.M., Koutedakis, Y., 2010. Biological evidence
for the acute health effects of secondhand smoke exposure. American Journal of Physiology-
Lung Cellular and Molecular Physiology 298, L3–L12.
https://doi.org/10.1152/ajplung.00215.2009
Forder, J.A., 2014. Simply Scan—Optical Methods for Elemental Carbon Measurement in Diesel Exhaust
Particulate. Ann Occup Hyg 58, 889–898. https://doi.org/10.1093/annhyg/meu037
Fuentes-Leonarte, V., Estarlich, M., Ballester, F., Murcia, M., Esplugues, A., Aurrekoetxea, J.J.,
Basterrechea, M., Fernández-Somoano, A., Morales, E., Gascón, M., Tardón, A., Rebagliato, M.,
2015. Pre- and postnatal exposure to tobacco smoke and respiratory outcomes during the first
year. Indoor Air 25, 4–12. https://doi.org/10.1111/ina.12128
Gray, S.C., Edwards, S.E., Miranda, M.L., 2010. Assessing Exposure Metrics for PM and Birthweight
Models. J Expo Sci Environ Epidemiol 20, 469–477. https://doi.org/10.1038/jes.2009.52
Habre, R., Moshier, E., Castro, W., Nath, A., Grunin, A., Rohr, A., Godbold, J., Schachter, N., Kattan,
M., Coull, B., Koutrakis, P., 2014. The effects of PM 2.5 and its components from indoor and
outdoor sources on cough and wheeze symptoms in asthmatic children. Journal of Exposure
Science & Environmental Epidemiology 24, 380–387. https://doi.org/10.1038/jes.2014.21
Hawsawi, A.M., Bryant, L.O., Goodfellow, L.T., 2015. Association Between Exposure to Secondhand
Smoke During Pregnancy and Low Birthweight: A Narrative Review. Respiratory Care 60, 135–
140. https://doi.org/10.4187/respcare.02798
Hoyt, A.T., Canfield, M.A., Romitti, P.A., Botto, L.D., Anderka, M.T., Krikov, S.V., Feldkamp, M.L.,
2018. Does Maternal Exposure to Secondhand Tobacco Smoke During Pregnancy Increase the
Risk for Preterm or Small-for-Gestational Age Birth? Matern Child Health J 22, 1418–1429.
https://doi.org/10.1007/s10995-018-2522-1
Hsieh, S.J., Ware, L.B., Eisner, M.D., Yu, L., Jacob, P., Havel, C., Goniewicz, M.L., Matthay, M.A.,
Benowitz, N.L., Calfee, C.S., 2011. Biomarkers increase detection of active smoking and
secondhand smoke exposure in critically ill patients. Crit Care Med 39, 40–45.
https://doi.org/10.1097/CCM.0b013e3181fa4196
Jaakkola, M.S., Jaakkola, J.J., 1997. Assessment of exposure to environmental tobacco smoke. European
Respiratory Journal 10, 2384–2397.
Jacob, P., Benowitz, N.L., Destaillats, H., Gundel, L., Hang, B., Martins-Green, M., Matt, G.E., Quintana,
P.J.E., Samet, J.M., Schick, S.F., Talbot, P., Aquilina, N.J., Hovell, M.F., Mao, J.-H., Whitehead,

125
T.P., 2017. Thirdhand Smoke: New Evidence, Challenges, and Future Directions. Chem Res
Toxicol 30, 270–294. https://doi.org/10.1021/acs.chemrestox.6b00343
Jedrychowski, W., Perera, F., Mrozek-Budzyn, D., Mroz, E., Flak, E., Spengler, J.D., Edwards, S., Jacek,
R., Kaim, I., Skolicki, Z., 2009. Gender differences in fetal growth of newborns exposed
prenatally to airborne fine particulate matter. Environmental Research 109, 447–456.
https://doi.org/10.1016/j.envres.2009.01.009
Jenkins, R.A., Counts, R.W., 1999. Personal exposure to environmental tobacco smoke: salivary cotinine,
airborne nicotine, and nonsmoker misclassification. J Expo Anal Environ Epidemiol 9, 352–363.
https://doi.org/10.1038/sj.jea.7500036
Jenkins, R.A., Maskarinec, M.P., Counts, R.W., Caton, J.E., Tomkins, B., Ilgner, R.H., 2001.
Environmental tobacco smoke in an unrestricted smoking workplace: area and personal exposure
monitoring. J Expo Sci Environ Epidemiol 11, 369–380. https://doi.org/10.1038/sj.jea.7500178
Kaufman, F.L., Kharrazi, M., Delorenze, G.N., Eskenazi, B., Bernert, J.T., 2002. Estimation of
environmental tobacco smoke exposure during pregnancy using a single question on household
smokers versus serum cotinine. J Expo Sci Environ Epidemiol 12, 286–295.
https://doi.org/10.1038/sj.jea.7500224
Kaur, J., Prasad, V.M., 2011. Air Nicotine Monitoring for Second Hand Smoke Exposure in Public Places
in India. Indian J Community Med 36, 98–103. https://doi.org/10.4103/0970-0218.84126
Kim, S., Apelberg, B.J., Avila-Tang, E., Hepp, L., Yun, D., Samet, J.M., Breysse, P.N., 2014. Utility and
Cutoff Value of Hair Nicotine as a Biomarker of Long-Term Tobacco Smoke Exposure,
Compared to Salivary Cotinine. Int J Environ Res Public Health 11, 8368–8382.
https://doi.org/10.3390/ijerph110808368
Kraev, T., Adamkiewicz, G., Hammond, S., Spengler, J., 2009. Indoor concentrations of nicotine in low-
income, multi-unit housing: associations with smoking behaviours and housing characteristics.
Tob Control 18, 438–444. https://doi.org/10.1136/tc.2009.029728
Lawless, P., 2004. Multiwavelength absorbance of filter deposits for determination of environmental
tobacco smoke and black carbon. Atmospheric Environment 38, 3373–3383.
https://doi.org/10.1016/j.atmosenv.2004.03.038
Leaderer, B.P., Hammond, S.K., 1991. Evaluation of vapor-phase nicotine and respirable suspended
particle mass as markers for environmental tobacco smoke. Environ. Sci. Technol. 25, 770–777.
https://doi.org/10.1021/es00016a023
Levine, H., Berman, T., Goldsmith, R., Göen, T., Spungen, J., Novack, L., Amitai, Y., Shohat, T., Grotto,
I., 2013. Exposure to tobacco smoke based on urinary cotinine levels among Israeli smoking and
nonsmoking adults: a cross-sectional analysis of the first Israeli human biomonitoring study.
BMC Public Health 13, 1241. https://doi.org/10.1186/1471-2458-13-1241
Martínez-Sánchez, J.M., González-Marrón, A., Martín-Sánchez, J.C., Sureda, X., Fu, M., Pérez-Ortuño,
R., Lidón-Moyano, C., Galán, I., Pascual, J.A., Fernández, E., 2018. Validity of self-reported
intensity of exposure to second-hand smoke at home against environmental and personal markers.
Gaceta Sanitaria 32, 393–395. https://doi.org/10.1016/j.gaceta.2017.08.002
Max, W., Sung, H.-Y., Shi, Y., 2009. Who Is Exposed to Secondhand Smoke? Self-Reported and Serum
Cotinine Measured Exposure in the U.S., 1999–2006. Int J Environ Res Public Health 6, 1633–
1648. https://doi.org/10.3390/ijerph6051633
Mourino, N., Pérez-Ríos, M., Santiago-Pérez, M.I., Lanphear, B., Yolton, K., Braun, J.M., 2021.
Secondhand tobacco smoke exposure among children under 5 years old: questionnaires versus
cotinine biomarkers: a cohort study. BMJ Open 11, e044829. https://doi.org/10.1136/bmjopen-
2020-044829
Muramatsu, M., Umemura, S., Okada, T., Tomita, H., 1984. Estimation of personal exposure to tobacco
smoke with a newly developed nicotine personal monitor. Environ Res 35, 218–227.
https://doi.org/10.1016/0013-9351(84)90130-0

126
Niu, Z., Xie, C., Wen, X., Tian, F., Yuan, S., Jia, D., Chen, W.-Q., 2016. Potential pathways by which
maternal second-hand smoke exposure during pregnancy causes full-term low birth weight. Sci
Rep 6, 24987. https://doi.org/10.1038/srep24987
Oberg, M., Jaakkola, M.S., Woodward, A., Peruga, A., Prüss-Ustün, A., 2011. Worldwide burden of
disease from exposure to second-hand smoke: a retrospective analysis of data from 192 countries.
Lancet 377, 139–146. https://doi.org/10.1016/S0140-6736(10)61388-8
O’Connor, T.Z., Holford, T.R., Leaderer, B.P., Hammond, S.K., Bracken, M.B., 1995. Measurement of
Exposure to Environmental Tobacco Smoke in Pregnant Women. American Journal of
Epidemiology 142, 1315–1321. https://doi.org/10.1093/oxfordjournals.aje.a117599
OEHHA, C.O. of E.H.H.A., 2006. Environmental Tobacco Smoke As A Toxic Air Contaminant To
Which Infants And Children May Be Especially Susceptible.
Park, H., Cho, S., Lee, C., 2019. Second hand smoke exposure in workplace by job status and
occupations. Ann Occup Environ Med 31, 3. https://doi.org/10.1186/s40557-019-0282-z
Pérez-Ríos, M., Schiaffino, A., López, M.J., Nebot, M., Galán, I., Fu, M., Martínez-Sánchez, J.M.,
Moncada, A., Montes, A., Ariza, C., Fernández, E., 2013. Questionnaire-based second-hand
smoke assessment in adults. European Journal of Public Health 23, 763–767.
https://doi.org/10.1093/eurpub/cks069
Pirkle, J.L., Flegal, K.M., Bernert, J.T., Brody, D.J., Etzel, R.A., Maurer, K.R., 1996. Exposure of the US
Population to Environmental Tobacco Smoke: The Third National Health and Nutrition
Examination Survey, 1988 to 1991. JAMA 275, 1233–1240.
https://doi.org/10.1001/jama.1996.03530400021033
Pogodina, C., Brunner Huber, L.R., Racine, E.F., Platonova, E., 2009. Smoke-Free Homes for Smoke-
Free Babies: The Role of Residential Environmental Tobacco Smoke on Low Birth Weight. J
Community Health 34, 376–382. https://doi.org/10.1007/s10900-009-9169-1
Price, P.N., Shehabi, A., Chan, R.W., Gadgil, A.J., 2006. Indoor-Outdoor Air Leakage of Apartments and
Commercial Buildings (No. LBNL-60682, 1010623). https://doi.org/10.2172/1010623
Primack, B.A., Bost, J.E., Land, S.R., Fine, M.J., 2007. Volume of Tobacco Advertising in African
American Markets: Systematic Review and Meta-Analysis. Public Health Rep 122, 607–615.
Protano, C., Cammalleri, V., Antonucci, A., Ungureanu, A.S., Santilli, F., Martellucci, S., Mattei, V.,
Vitali, M., 2019. Further Insights on Predictors of Environmental Tobacco Smoke Exposure
during the Pediatric Age. Int J Environ Res Public Health 16, 4062.
https://doi.org/10.3390/ijerph16214062
Rasmussen, J.L., Dunlap, W.P., 1991. Dealing with Nonnormal Data: Parametric Analysis of
Transformed Data vs Nonparametric Analysis. Educational and Psychological Measurement 51,
809–820. https://doi.org/10.1177/001316449105100402
Ratschen, E., Thorley, R., Jones, L., Opazo Breton, M., Cook, J., McNeill, A., Britton, J., Coleman, T.,
Lewis, S., 2018. A randomised controlled trial of a complex intervention to reduce children’s
exposure to secondhand smoke in the home. Tob Control 27, 155–162.
https://doi.org/10.1136/tobaccocontrol-2016-053279
Rebagliato, M., Bol, F., Florey, C. du V., Jarvis, M.J., Pérez-Hoyos, S., Hern, I., Avi, M.J., 1998.
Variations in cotinine levels in smokers during and after pregnancy. American Journal of
Obstetrics & Gynecology 178, 568–571. https://doi.org/10.1016/S0002-9378(98)70440-5
Ritz, B., Wilhelm, M., Hoggatt, K.J., Ghosh, J.K.C., 2007. Ambient Air Pollution and Preterm Birth in
the Environment and Pregnancy Outcomes Study at the University of California, Los Angeles.
American Journal of Epidemiology 166, 1045–1052. https://doi.org/10.1093/aje/kwm181
Ritz, Beate, Wilhelm, M., Hoggatt, K.J., Ghosh, J.K.C., 2007. Ambient air pollution and preterm birth in
the environment and pregnancy outcomes study at the University of California, Los Angeles. Am
J Epidemiol 166, 1045–1052. https://doi.org/10.1093/aje/kwm181
Sarnat, J.A., Koutrakis, P., Suh, H.H., 2000. Assessing the relationship between personal particulate and
gaseous exposures of senior citizens living in Baltimore, MD. J Air Waste Manag Assoc 50,
1184–1198. https://doi.org/10.1080/10473289.2000.10464165

127
Semple, S., Apsley, A., Azmina Ibrahim, T., Turner, S.W., Cherrie, J.W., 2015. Fine particulate matter
concentrations in smoking households: just how much secondhand smoke do you breathe in if
you live with a smoker who smokes indoors? Tob Control 24, e205-211.
https://doi.org/10.1136/tobaccocontrol-2014-051635
Shastri, S.S., Talluri, R., Shete, S., 2021. Disparities in Secondhand Smoke Exposure in the United States:
National Health and Nutrition Examination Survey 2011-2018. JAMA Internal Medicine 181,
134–137. https://doi.org/10.1001/jamainternmed.2020.3975
Sloan, C.D., Weber, F.X., Bradshaw, R.K., Philipp, T.J., Barber, W.B., Palmer, V.L., Graul, R.J., Tuttle,
S.C., Chartier, R.T., Johnston, J.D., 2017. Elemental analysis of infant airborne particulate
exposures. J Expo Sci Environ Epidemiol 27, 526–534. https://doi.org/10.1038/jes.2016.77
Suh, H.H., Zanobetti, A., 2010. Exposure Error Masks The Relationship Between Traffic-Related Air
Pollution and Heart Rate Variability (HRV). J Occup Environ Med 52, 685–692.
https://doi.org/10.1097/JOM.0b013e3181e8071f
Thornburg, J., Halchenko, Y., McCombs, M., Siripanichgon, N., Dowell, E., Cho, S.-H., Egner, J.,
Sayarath, V., Karagas, M.R., 2021. Children’s Particulate Matter Exposure Characterization as
Part of the New Hampshire Birth Cohort Study. Int J Environ Res Public Health 18, 12109.
https://doi.org/10.3390/ijerph182212109
Travers, M.J., Vogl, L., 2015. FLORENCE, ALABAMA INDOOR AIR QUALITY MONITORING
STUDY 12.
Tsai, J., 2018. Exposure to Secondhand Smoke Among Nonsmokers — United States, 1988–2014.
MMWR Morb Mortal Wkly Rep 67. https://doi.org/10.15585/mmwr.mm6748a3
United States Surgeon General, 2014. The Health Consequences of Smoking -- 50 Years of progress: A
Report of the Surgeon General: (510072014-001). https://doi.org/10.1037/e510072014-001
U.S. Department of Health and Human Services, 2006. The Health Consequences of Involuntary
Exposure to Tobacco Smoke: A Report of the Surgeon General, Publications and Reports of the
Surgeon General. Centers for Disease Control and Prevention (US), Atlanta (GA).
US EPA, O., 2014. Secondhand Smoke and Smoke-free Homes [WWW Document]. URL
https://www.epa.gov/indoor-air-quality-iaq/secondhand-smoke-and-smoke-free-homes (accessed
2.27.22).
Vartiainen, E., Seppälä, T., Lillsunde, P., Puska, P., 2002. Validation of self reported smoking by serum
cotinine measurement in a community-based study. Journal of Epidemiology & Community
Health 56, 167–170. https://doi.org/10.1136/jech.56.3.167
Wang, C.-P., Ma, S.J., Xu, X.F., Wang, J.-F., Mei, C.Z., Yang, G.-H., 2009. The prevalence of household
second-hand smoke exposure and its correlated factors in six counties of China. Tob Control 18,
121–126. https://doi.org/10.1136/tc.2008.024836
Weisskopf, M.G., Webster, T.F., 2017. Trade-offs of personal vs. more proxy exposure measures in
environmental epidemiology. Epidemiology 28, 635–643.
https://doi.org/10.1097/EDE.0000000000000686
Williams, R., Collier, A., Lewtas, J., 1993. Environmental Tobacco Smoke Exposure of Young Children
as Assessed Using a Passive Diffusion Device for Nicotine. IED 2, 98–104.
https://doi.org/10.1159/000463231
Wilson, S.E., Kahn, R.S., Khoury, J., Lanphear, B.P., 2005. Racial Differences in Exposure to
Environmental Tobacco Smoke among Children. Environ Health Perspect 113, 362–367.
https://doi.org/10.1289/ehp.7379
Wipfli, H., Avila-Tang, E., Navas-Acien, A., Kim, S., Onicescu, G., Yuan, J., Breysse, P., Samet, J.M.,
2008. Secondhand Smoke Exposure Among Women and Children: Evidence From 31 Countries.
Am J Public Health 98, 672–679. https://doi.org/10.2105/AJPH.2007.126631
Woodruff, S.I., Conway, T.L., Edwards, C.C., Hovell, M.F., 2003. Acceptability and validity of hair
collection from Latino children to assess exposure to environmental tobacco smoke. Nicotine &
Tobacco Research 5, 375–385. https://doi.org/10.1080/1462220031000094187

128
Wynne, O., Bonevski, B., 2018. Developments in the Research Base on Reducing Exposure to Second-
Hand Smoke. Int J Environ Res Public Health 15, 1873. https://doi.org/10.3390/ijerph15091873
Xiao, X., Li, Y., Song, X., Xu, Q., Yang, S., Wu, J., Seto, E., 2018. Discrepancy between Self-Reported
and Urine Cotinine-Verified Environmental Tobacco Smoke Exposure among Rural Pregnant
Women in China. Int J Environ Res Public Health 15, 1499.
https://doi.org/10.3390/ijerph15071499
Yan, B., Kennedy, D., Miller, R.L., Cowin, J.P., Jung, K., Perzanowski, M., Balletta, M., Perera, F.P.,
Kinney, P.L., Chillrud, S.N., 2011. Validating a nondestructive optical method for apportioning
colored particulate matter into black carbon and additional components. Atmos Environ 45,
7478–7486. https://doi.org/10.1016/j.atmosenv.2011.01.044
Zhang, T., Chillrud, S.N., Yang, Q., Pitiranggon, M., Ross, J., Perera, F., Ji, J., Spira, A., Breysse, P.N.,
Rodes, C.E., Miller, R., Yan, B., 2020. Characterizing peak exposure of secondhand smoke using
a real-time PM2.5 monitor. Indoor Air 30, 98–107. https://doi.org/10.1111/ina.12611
Zheng, Z., Xie, G., Yang, T., Qin, J., 2019. Congenital malformations are associated with secondhand
smoke among nonsmoking women: A meta-analysis. Birth 46, 222–233.
https://doi.org/10.1111/birt.12401

129

CHAPTER 5
SUMMARY AND FUTURE RESEARCH DIRECTIONS
The in-utero period is an important developmental period that can have lifelong impacts on health.
This vulnerable period is susceptible to insult, possibly through the creation of a hostile intrauterine
environment (Barker, 2004; Clemente et al., 2016; Z. Li et al., 2019), which can impact fetal development,
leading to enduring adverse health conditions throughout life, including child- and adulthood diseases (Mi
et al., 2017; Umer et al., 2020; Upadhyay et al., 2019; Vilanova et al., 2019). Studies have reported vast
differences in exposure to environmental exposures, including air pollution, by race/ethnicity and education
attainment (Bell & Ebisu, 2012). In addition to the increased burden of environmental exposures, Hispanic
and non-Hispanic Black populations face the greatest burden of fetal development issues, including low
birth weight (Womack et al., 2018). These adverse environmental and health burdens may be compounded
by additional factors including, limited access to health care and economic disparities (Mahajan et al., 2021;
Wang et al., 2013). Furthermore, these populations have be systematically underserved in health research,
highlighting the pressing need for research to be conducted in health disparities populations.
Previous research into the effect of PM 2.5 on birthweight during the in-utero period have
predominantly found a weak to moderate association between outdoor PM 2.5 and birthweight (X. Li et al.,
2017; Stieb et al., 2012), particularly in the 3
rd
trimester where most fetal weight gains occurs (Kiserud et
al., 2018). However, inconsistency remains in the literature, likely due to estimating individual exposure at
the residential level using models that typically incorporate ambient monitoring, remote sensing, and/or
geospatial data (Dadvand et al., 2013; Gray et al., 2010; Harris et al., 2014), which lead to biased effect
estimates and attenuated statistical power (Carroll, 2005; Zeger et al., 2000).
People spend the vast majority of their time inside, and their “true” personal exposure to PM 2.5 of
outdoor origin a result of the infiltration efficiency of PM 2.5 indoors and time-activity patterns, most
accurately captured by personal monitoring (Gray et al., 2011). Additionally, outdoor estimates do not
measure a person’s total personal PM 2.5 exposure, which also includes indoor sources. A prior study found

130
that exposure to personal PM 2.5 was negatively associated with birthweight (Jedrychowski et al., 2009);
however, this was in a racially homogenous population. Additionally, there are large differences in PM 2.5
sources and components across regions (Basu et al., 2014; Bell et al., 2007; Ng et al., 2017), which may
differ in toxicity on birthweight (Bell et al., 2012; Laurent et al., 2016; Sun et al., 2016). Finally, secondhand
smoke (SHS) is a major contributing source of personal PM 2.5; however, it is most commonly measured by
self-report questions, which the literature highlights as being highly heterogenous in terms of question type
and just 17.9% (~35% in pregnancy studies) of studies validated the questions against objective measures
(Pérez-Ríos et al., 2013).
Therefore, my dissertation aimed to address several critical gaps in knowledge around PM 2.5
sources and provide contributions to the literature for the effect of personal PM 2.5 on birthweight, in the
Maternal And Developmental Risks from Environmental and Social stressors (MADRES) pregnancy
cohort. First, this study addressed the current gap in knowledge regarding the effect of total personal PM 2.5
as a whole or whether it is more impacted by indoor or outdoor sources on birthweight in a health disparities
population. The effect of total personal PM 2.5 on birthweight was explored using “gold standard” personal
exposure monitoring data, which also allowed for the characterization of total personal PM 2.5 exposure
during pregnancy in a health disparities population in urban Los Angeles, CA (Study 1). Additionally, the
effect of personal PM 2.5 more impacted by indoor or outdoor sources was investigated to evaluate whether
this changed the overall effect. Furthermore, it also allowed for a look at the impact of different sources
(indoor vs. outdoor) on PM 2.5 on birthweight. Next, by using chemical speciation data from the personal
monitoring study and source apportionment approaches (conducted by Dr. Yan Xu), this dissertation sought
to determine the association between personal PM 2.5 sources and their respective high-loading components
(elements or carbon fractions). This provides evidence for whether PM 2.5 sources or individual components
were more averse to fetal growth using less biased exposure assessment (Study 2). Finally, optical
absorbance techniques on the PM 2.5 mass samples would allow for the evaluation of measured airborne
environmental tobacco smoke (ETS) vs. self-report SHS assessment from different study time-points and

131
question wording (Study 3). This would contribute to the literature of how well self-report SHS assessment
indicates SHS exposure in pregnancy studies.
The main findings of this dissertation were:
1. Total personal exposure to PM 2.5 was not significantly associated with birthweight in this health
disparities population. However, the effect of personal PM 2.5 differed by home type (multi-unit
housing vs. single-family homes), candle/incense smoke, and greater outdoor source contributions
to personal PM 2.5, highlighting that specific indoor sources or personal PM 2.5 more impacted by
outdoor sources may have more adverse effects on birthweight. Personal PM 2.5 was weakly
positively associated with outdoor PM 2.5 over the same 48hr sampling period and 3
rd
trimester
average, highlighting that outdoor PM 2.5 does not account for a person’s full exposure.
2. Using a chemical speciation-based approach, this dissertation found that the effect of major
contributing PM 2.5 sources on birthweight was heterogenous, with fresh sea salt and aged sea salt
seemingly having the most negative effect on birthweight, followed by crustal and SHS.
Additionally, the effect of the crustal source on birthweight differed by infant sex, with girls having
a positive effect vs a negative effect for boys. A similar pattern was observed with fuel oil, where
girls witnessed a positive effect on birthweight, compared to negative to boys. This suggests that
not only does the effect of major contributing sources of PM 2.5 differ in its effect on birthweight,
but it may also vary across infant sex.
3. Overall, self-reported SHS questions were not associated with measured airborne ETS but the
question regarding the number of smokers nearby was more correlated with measured ETS than
questions asking about the presence of a smoker or duration of smoke exposure. We found
underreporting of exposure was common across time-points and wording types, supporting
concerns of attenuated health effects for self-reported SHS assessment studies. Measured airborne
ETS was significantly higher in participants with high school or less, while more window use was
also significantly associated with higher ETS. Finally, we found marginally significant lower ETS

132
by home type with single-family home participants observing the lowest ETS exposure compared
to multi-unit housing.
Conclusions and Implications
This dissertation found evidence that major sources of personal PM 2.5 differ in their effect on
birthweight within a health disparities population in Los Angeles, CA, which provides a rationale for more
targeted air pollution regulatory interventions and pregnancy behavioral changes to improve birth
outcomes. Firstly, in Study 1 of this dissertation, total personal PM 2.5 exposure was not associated with
birthweight in the 3
rd
trimester; however, there was evidence that candle and/or incense smoke exposure,
and greater outdoor source contributions to personal PM 2.5 were more strongly associated with lower
birthweight. Considering preterm and LBW infants in the United States account for around 50% of
hospitalization costs (Russell et al., 2007), behavioral interventions to reduce overall risk may have health
and economic cost benefits. The evidence found here, namely, how candle/incense modified the effect of
personal PM 2.5 on birthweight, and greater time spent outside observed a more negative association of
personal PM 2.5 on birthweight, provides data to support both regulatory and behavioral interventions as
potential ways to reduce risk of having a LBW infant, such as further reducing ambient PM 2.5 limits through
legislation, and avoiding candle or incense smoke during pregnancy.
Next, we showed that the effect of personal PM 2.5 differed by the amount of time spent with open
windows in the home, air conditioner (AC) use, and the type of home participants lived in. This contribution
to the literature highlights that home characteristics are potential avenues for exposure but also for
interventions in several ways. Firstly, education related to home use, such as keeping windows closed, and
if possible, using AC units, can be provided to pregnant women, which may potentially reduce the risk of
LBW. Next, suggestions to stay away from candle/incense exposure during pregnancy may also be a useful
tool for reducing risk.
Importantly, PM 2.5 is not simply a homogenous outdoor exposure, but rather a mixture of pollutants
than vary in their toxicity, therefore, we evaluated major contributing sources of personal PM 2.5 using a
chemical approach to determine which sources place pregnant mothers at greater risk of decreased

133
birthweight. Fresh and aged sea salt had the most negative effect on birthweight, followed by crustal and
SHS. There was evidence that individual high-loading components within a given source may also be
driving the effect of the most toxic sources. For example, within fresh and aged sea salt, sodium (Na) and
magnesium (Mg) were significantly negatively associated with birthweight. This may be a novel finding
that these pollutants were leading to adverse birthweight or that these components are correlated with other
exposures, such as pollutants formed secondary processes. This evidence provided in this dissertation,
combined with the current lack of tailored regulation for specific sources and components of PM 2.5,
highlights the need for regulation beyond ambient air pollution. Even with the decrease in ambient PM 2.5,
adverse health effects remain.
In our analysis, home type not only modified the effect of personal PM 2.5 on birthweight (Study 1)
but also correlated with highest geometric mean concentration of ETS (Study 3), highlighting that leakage
and infiltration in multi-unit housing buildings may impact which sources of PM 2.5 participants are exposed
to. This study presented further evidence that smoking bans should be implemented, particularly in
multiunit housing, due to the infiltration of SHS into neighboring units, and especially in low-income health
disparities populations that face a greater burden of environmental exposures (Bell & Ebisu, 2012). While
regulation related to reduced or banned indoor smoking in many settings, including bars, restaurants, and
worksites, have been successful in reducing smoke exposure, there is currently no national or statewide ban
on smoking in one’s own private residence (CDC, 2022). The Office of Public and Indian Housing (PIH),
an agency of the U.S. Department of House and Urban Development (HUD) passed a ruling (CFR §
965.653; effective date February 3, 2017) that implements a smoke-free policy for public housing. Several
municipalities have enacted local laws banning smoking in multiunit homes, with implementation
challenges that still need to be addressed (Yerger et al., 2014); however, total bans have been shown to be
successful at reducing exposure to SHS (Zablocki et al., 2014).
This current work adds to the SHS measurement error literature, by providing one of the first
accounts of measured airborne ETS exposure assessment using personal monitoring data in a health
disparities pregnancy study, and by providing a comparison of said metric to self-reported SHS questions.

134
Our results indicate that self-reported SHS exposure was not associated with measured airborne ETS;
however, the number of smokers nearby question seemed to correlate more with measure ETS than
questions that ask about the simple presence of a smoker, or the amount of time surrounded by smokers. A
similar result showing the benefits of questions regarding the number of smokers nearby to determine SHS
exposure was found by others using serum cotinine (Kaufman et al., 2002). Since question type and
validation of self-reported SHS assessment remains low in health analyses (Pérez-Ríos et al., 2013), this
work provides further evidence that standardized methods are needed to provide more accurate self-reported
exposure assessment but also to allow for comparisons and combining SHS metrics across cohorts.
Future Studies and Research Directions
This dissertation presented some of the only evidence for the effect of personal PM 2.5 sources and
components on birthweight. However, birthweight is just one avenue of fetal growth, and a natural next
step would be to evaluate the effect of personal PM 2.5 sources and components on other fetal growth/health
outcomes, including, head circumference, biparietal diameter, abdominal circumference, and femur length.
This may elicit biomechanisms underlying the effects of PM 2.5 on fetal health by providing evidence for
which parts of fetal development is being affected. This brings me to another future and closely linked study
surrounding the effects of personal PM 2.5 on fetal adiposity. Birthweight, while being a useful and widely
used metric for fetal health, differences in adiposity may exist, either acting independently or in parallel
with alterations in fetal size. This is especially important considering there has been a link between
birthweight and later life obesity, and chemical exposures, including PM 2.5 are thought to act as obesogens
(Heindel & Schug, 2014).
Many fetal growth and fetal health outcomes are highly correlated with each other. For example,
gestational age and preterm births are highly correlated with birthweight. Most studies assess individual
outcomes separately, often running several independent models with different fetal growth metrics as the
outcome. While this is an effective way of establishing potential relationships, it can run into problems
when picking covariates. In the case of birthweight and other fetal growth metrics, gestational age is
regularly argued as being a confounder; however, it may also be a mediator in the association with an

135
environmental exposure, leading to bias (Wilcox et al., 2011). One potential methodology worth looking at
would be a multivariate probit model, which is used to estimate the effect of several binary outcomes jointly
(CHIB & GREENBERG, 1998). An example of this would be modeling both LBW and preterm birth.
Therefore, a future study would look at several of these highly correlated birth outcomes together.
There are several approaches to SHS assessment, from self-report, measurements or airborne
chemicals, and biomarkers. While biomarkers are the “gold standard” for internal dose assessment (as
opposed to personal monitoring being considered the gold standard for external exposure assessment,
before the chemical enters the body), logistical constraints, differences in how biomarkers react within the
body during pregnancy, duration of exposure, and retroactive exposures, means there is a need or benefit
to having several approaches to SHS assessment. Therefore, I would be interested in assessing several
different approaches to SHS assessment for the same group of participants. Typical studies have looked at
self-report vs. one objective measure. However, a study assessing SHS exposure across several different
assessment methods, from self-report to airborne (both stationary and personal), and biomarkers would
provide valuable insight into whether the differing approaches are concordant (or not) with one another. It
would also provide a method for assessing airborne methods vs biomarker methods, which may provide
evidence to the relative amount of exposure that a person experiences in their breathing zone that actually
gets into the body. Additionally, while several cut-points have been used to indicate SHS exposure (Pérez-
Ríos et al., 2013), there is a need to establish more standardized measures. By evaluating several approaches
of exposure assessment, it may be possible to contribute thresholds across metrics in pregnancy studies that
indicate SHS exposure.
CONCLUSIONS
This dissertation found evidence that the effect of personal PM 2.5 on birthweight depends on the
source and chemical composition of PM 2.5 mass with fresh sea salt, aged sea salt sources, and Mg, Na, Cl,
ETS, and BrC having the most adverse effect on birthweight. There was also evidence that personal PM 2.5
sources may be modified by infant sex with boys observing a more negative effect on birthweight compared
to girls for crustal and fuel oil. Additionally, there was no evidence for an association between measure

136
airborne ETS and self-report SHS exposure; however, there was evidence that a question asking about the
number of smokers nearby was more correlated with measured ETS, compared to the questions asking
about the presence of a smoke or the duration of smoke exposure. Further research is needed to investigate
the effect of personal PM 2.5 sources on other fetal growth outcomes.

137
REFERENCES
Barker, D. J. P. (2004). The developmental origins of chronic adult disease. Acta Paediatrica (Oslo,
Norway: 1992). Supplement, 93(446), 26–33. https://doi.org/10.1111/j.1651-2227.2004.tb00236.x
Basu, R., Harris, M., Sie, L., Malig, B., Broadwin, R., & Green, R. (2014). Effects of fine particulate
matter and its constituents on low birth weight among full-term infants in California.
Environmental Research, 128, 42–51. https://doi.org/10.1016/j.envres.2013.10.008
Bell, M. L., Belanger, K., Ebisu, K., Gent, J. F., & Leaderer, B. P. (2012). Relationship Between Birth
Weight and Exposure to Airborne Fine Particulate Potassium and Titanium During Gestation.
Environmental Research, 117, 83–89. https://doi.org/10.1016/j.envres.2012.05.004
Bell, M. L., Dominici, F., Ebisu, K., Zeger, S. L., & Samet, J. M. (2007). Spatial and Temporal Variation
in PM2.5 Chemical Composition in the United States for Health Effects Studies. Environmental
Health Perspectives, 115(7), 989–995. https://doi.org/10.1289/ehp.9621
Bell, M. L., & Ebisu, K. (2012). Environmental Inequality in Exposures to Airborne Particulate Matter
Components in the United States. Environmental Health Perspectives, 120(12), 1699–1704.
https://doi.org/10.1289/ehp.1205201
Carroll, R. J. (2005). Measurement Error in Epidemiologic Studies. In Encyclopedia of Biostatistics.
American Cancer Society. https://doi.org/10.1002/0470011815.b2a03082
CDC. (2022). STATE System Smokefree Indoor Air Fact Sheet.
https://www.cdc.gov/statesystem/factsheets/sfia/SmokeFreeIndoorAir.html
CHIB, S., & GREENBERG, E. (1998). Analysis of multivariate probit models. Biometrika, 85(2), 347–
361. https://doi.org/10.1093/biomet/85.2.347
Clemente, D. B. P., Casas, M., Vilahur, N., Begiristain, H., Bustamante, M., Carsin, A.-E., Fernández, M.
F., Fierens, F., Gyselaers, W., Iñiguez, C., Janssen, B. G., Lefebvre, W., Llop, S., Olea, N.,
Pedersen, M., Pieters, N., Santa Marina, L., Souto, A., Tardón, A., … Nawrot, T. S. (2016).
Prenatal Ambient Air Pollution, Placental Mitochondrial DNA Content, and Birth Weight in the
INMA (Spain) and ENVIRONAGE (Belgium) Birth Cohorts. Environmental Health
Perspectives, 124(5), 659–665. https://doi.org/10.1289/ehp.1408981
Dadvand, P., Parker, J., Bell, M. L., Bonzini, M., Brauer, M., Darrow, L. A., Gehring, U., Glinianaia, S.
V., Gouveia, N., Ha, E., Leem, J. H., van den Hooven, E. H., Jalaludin, B., Jesdale, B. M.,
Lepeule, J., Morello-Frosch, R., Morgan, G. G., Pesatori, A. C., Pierik, F. H., … Woodruff, T. J.
(2013). Maternal Exposure to Particulate Air Pollution and Term Birth Weight: A Multi-Country
Evaluation of Effect and Heterogeneity. Environmental Health Perspectives, 121(3), 267–373.
https://doi.org/10.1289/ehp.1205575
Gray, S. C., Edwards, S. E., & Miranda, M. L. (2010). Assessing Exposure Metrics for PM and
Birthweight Models. Journal of Exposure Science & Environmental Epidemiology, 20(5), 469–
477. https://doi.org/10.1038/jes.2009.52
Gray, S. C., Gelfand, A. E., & Miranda, M. L. (2011). Hierarchical spatial modeling of uncertainty in air
pollution and birth weight study. Statistics in Medicine, 30(17), 2187–2198.
https://doi.org/10.1002/sim.4234
Harris, G., Thompson, W. D., Fitzgerald, E., & Wartenberg, D. (2014). The association of PM2.5 with
full term low birth weight at different spatial scales. Environmental Research, 134, 427–434.
https://doi.org/10.1016/j.envres.2014.05.034
Heindel, J. J., & Schug, T. T. (2014). The Obesogen Hypothesis: Current Status and Implications for
Human Health. Current Environmental Health Reports, 1(4), 333–340.
https://doi.org/10.1007/s40572-014-0026-8
Jedrychowski, W., Perera, F., Mrozek-Budzyn, D., Mroz, E., Flak, E., Spengler, J. D., Edwards, S., Jacek,
R., Kaim, I., & Skolicki, Z. (2009). Gender differences in fetal growth of newborns exposed
prenatally to airborne fine particulate matter. Environmental Research, 109(4), 447–456.
https://doi.org/10.1016/j.envres.2009.01.009

138
Kaufman, F. L., Kharrazi, M., Delorenze, G. N., Eskenazi, B., & Bernert, J. T. (2002). Estimation of
environmental tobacco smoke exposure during pregnancy using a single question on household
smokers versus serum cotinine. Journal of Exposure Science & Environmental Epidemiology,
12(4), 286–295. https://doi.org/10.1038/sj.jea.7500224
Kiserud, T., Benachi, A., Hecher, K., Perez, R. G., Carvalho, J., Piaggio, G., & Platt, L. D. (2018). The
World Health Organization fetal growth charts: Concept, findings, interpretation, and application.
American Journal of Obstetrics & Gynecology, 218(2), S619–S629.
https://doi.org/10.1016/j.ajog.2017.12.010
Laurent, O., Hu, J., Li, L., Kleeman, M. J., Bartell, S. M., Cockburn, M., Escobedo, L., & Wu, J. (2016).
Low birth weight and air pollution in California: Which sources and components drive the risk?
Environment International, 92–93, 471–477. https://doi.org/10.1016/j.envint.2016.04.034
Li, X., Huang, S., Jiao, A., Yang, X., Yun, J., Wang, Y., Xue, X., Chu, Y., Liu, F., Liu, Y., Ren, M.,
Chen, X., Li, N., Lu, Y., Mao, Z., Tian, L., & Xiang, H. (2017). Association between ambient
fine particulate matter and preterm birth or term low birth weight: An updated systematic review
and meta-analysis. Environmental Pollution, 227, 596–605.
https://doi.org/10.1016/j.envpol.2017.03.055
Li, Z., Tang, Y., Song, X., Lazar, L., Li, Z., & Zhao, J. (2019). Impact of ambient PM2.5 on adverse birth
outcome and potential molecular mechanism. Ecotoxicology and Environmental Safety, 169,
248–254. https://doi.org/10.1016/j.ecoenv.2018.10.109
Lurmann, F., Avol, E., & Gilliland, F. (2015). Emissions Reduction Policies and Recent Trends in
Southern California’s Ambient Air Quality. Journal of the Air & Waste Management Association
(1995), 65(3), 324–335. https://doi.org/10.1080/10962247.2014.991856
Mahajan, S., Caraballo, C., Lu, Y., Valero-Elizondo, J., Massey, D., Annapureddy, A. R., Roy, B., Riley,
C., Murugiah, K., Onuma, O., Nunez-Smith, M., Forman, H. P., Nasir, K., Herrin, J., &
Krumholz, H. M. (2021). Trends in Differences in Health Status and Health Care Access and
Affordability by Race and Ethnicity in the United States, 1999-2018. JAMA, 326(7), 637–648.
https://doi.org/10.1001/jama.2021.9907
Mi, D., Fang, H., Zhao, Y., & Zhong, L. (2017). Birth weight and type 2 diabetes: A meta-analysis.
Experimental and Therapeutic Medicine, 14(6), 5313–5320.
https://doi.org/10.3892/etm.2017.5234
Ng, C., Malig, B., Hasheminassab, S., Sioutas, C., Basu, R., & Ebisu, K. (2017). Source apportionment of
fine particulate matter and risk of term low birth weight in California: Exploring modification by
region and maternal characteristics. Science of The Total Environment, 605–606, 647–654.
https://doi.org/10.1016/j.scitotenv.2017.06.053
Pérez-Ríos, M., Schiaffino, A., López, M. J., Nebot, M., Galán, I., Fu, M., Martínez-Sánchez, J. M.,
Moncada, A., Montes, A., Ariza, C., & Fernández, E. (2013). Questionnaire-based second-hand
smoke assessment in adults. European Journal of Public Health, 23(5), 763–767.
https://doi.org/10.1093/eurpub/cks069
Russell, R. B., Green, N. S., Steiner, C. A., Meikle, S., Howse, J. L., Poschman, K., Dias, T., Potetz, L.,
Davidoff, M. J., Damus, K., & Petrini, J. R. (2007). Cost of hospitalization for preterm and low
birth weight infants in the United States. Pediatrics, 120(1), e1-9.
https://doi.org/10.1542/peds.2006-2386
Stieb, D. M., Chen, L., Eshoul, M., & Judek, S. (2012). Ambient air pollution, birth weight and preterm
birth: A systematic review and meta-analysis. Environmental Research, 117, 100–111.
https://doi.org/10.1016/j.envres.2012.05.007
Sun, X., Luo, X., Zhao, C., Zhang, B., Tao, J., Yang, Z., Ma, W., & Liu, T. (2016). The associations
between birth weight and exposure to fine particulate matter (PM2.5) and its chemical
constituents during pregnancy: A meta-analysis. Environmental Pollution, 211, 38–47.
https://doi.org/10.1016/j.envpol.2015.12.022
Umer, A., Hamilton, C., Cottrell, L., Giacobbi, P., Innes, K., Kelley, G. A., Neal, W., John, C., & Lilly,
C. (2020). Association between birth weight and childhood cardiovascular disease risk factors in

139
West Virginia. Journal of Developmental Origins of Health and Disease, 11(1), 86–95.
https://doi.org/10.1017/S204017441900045X
Upadhyay, R. P., Naik, G., Choudhary, T. S., Chowdhury, R., Taneja, S., Bhandari, N., Martines, J. C.,
Bahl, R., & Bhan, M. K. (2019). Cognitive and motor outcomes in children born low birth
weight: A systematic review and meta-analysis of studies from South Asia. BMC Pediatrics,
19(1), 35. https://doi.org/10.1186/s12887-019-1408-8
Vilanova, C. S., Hirakata, V. N., de Souza Buriol, V. C., Nunes, M., Goldani, M. Z., & da Silva, C. H.
(2019). The relationship between the different low birth weight strata of newborns with infant
mortality and the influence of the main health determinants in the extreme south of Brazil.
Population Health Metrics, 17(1), 15. https://doi.org/10.1186/s12963-019-0195-7
Wang, T.-F., Shi, L., Nie, X., & Zhu, J. (2013). Race/Ethnicity, insurance, income and access to care: The
influence of health status. International Journal for Equity in Health, 12(1), 29.
https://doi.org/10.1186/1475-9276-12-29
Wilcox, A. J., Weinberg, C. R., & Basso, O. (2011). On the Pitfalls of Adjusting for Gestational Age at
Birth. American Journal of Epidemiology, 174(9), 1062–1068. https://doi.org/10.1093/aje/kwr230
Womack, L. S., Rossen, L. M., & Martin, J. A. (2018, March 1). Singleton Low Birthweight Rates, by
Race and Hispanic Origin: United States, 2006–2016 [Center for Disease Control and Prevention
(CDC)]. National Center for Health Statistics.
https://www.cdc.gov/nchs/products/databriefs/db306.htm
Yerger, V. B., Battle, R. S., & Moore, R. S. (2014). Evaluating the Implementation Process of a Citywide
Smoke-Free Multiunit Housing Ordinance: Insights from Community Stakeholders. American
Journal of Public Health, 104(10), 1889–1891. https://doi.org/10.2105/AJPH.2014.302075
Zablocki, R. W., Edland, S. D., Myers, M. G., Strong, D. R., Hofstetter, C. R., & Al-Delaimy, W. K.
(2014). Smoking ban policies and their influence on smoking behaviors among current California
smokers: A population-based study. Preventive Medicine, 59, 73–78.
https://doi.org/10.1016/j.ypmed.2013.11.018
Zeger, S. L., Thomas, D., Dominici, F., Samet, J. M., Schwartz, J., Dockery, D., & Cohen, A. (2000).
Exposure measurement error in time-series studies of air pollution: Concepts and consequences.
Environmental Health Perspectives, 108(5), 419–426.

Abstract (if available)

Abstract Low birthweight (LBW) is an important birth health metric, with studies finding associations of LBW with several adverse health outcomes in both early and later life, including metabolic diseases, cardiovascular disease, cognitive impairment, and infant mortality. Previous studies generally support a weak to moderate association between outdoor air pollution and specifically particulate matter with aerodynamic diameter less than 2.5µm (PM2.5) and decreased birthweight. However, previous results are mixed, likely due to measurement error introduced by estimating personal exposure from ambient data, and from differences in the major contributing sources of PM2.5, which likely have different toxicities. Additionally, while outdoor PM2.5 is an important research question due to this criteria air pollutant being regulated and shown to have important population-wide adverse effects on health, personal PM2.5 exposure consists of outdoor and indoor PM2.5 sources as a result of individuals’ time-activity patterns (i.e., time spent indoors or outdoors), various behaviors, and infiltration of outdoor PM2.5 indoors. This total external exposure in the personal breathing zone is best captured by personal monitoring. Very few studies have been able to investigate the effects of personal PM2.5 exposure during pregnancy on birthweight and whether that differs by the major sources contributing to it, particularly in the 3rd trimester where most fetal weight gain occurs. This dissertation aimed at addressing the following three research questions to better understand the effect of in-utero exposure to personal PM2.5 and its major contributing sources on birthweight in the 3rd trimester within a health disparities population in Los Angeles, CA:
1) Does the effect of personal PM2.5 on birthweight vary based on whether it was mostly impacted by PM2.5 sources of indoor vs outdoor origin?
2) What is the effect of specific sources of personal PM2.5 on birthweight, as characterized by their chemical composition?
3) Is personal exposure to secondhand smoke (SHS), a major source of personal PM2.5 and a known risk factor for reductions in birthweight, adequately assessed through self-reported SHS questionnaires?
Within this dissertation, I present evidence that while total personal PM2.5 exposure was not significantly associated with birthweight, indoor versus outdoor origin, and more specific sources of personal PM2.5 appear to be adversely associated with birthweight. Home characteristics, including, more time with windows open in the home, not using air conditioning (AC), candle/incense smoke exposure, and more time spent outdoors, were associated with a more negative effect of personal PM2.5 on birthweight. This suggests that sources originating indoors versus outdoors and potentially different pathways may impact this relationship, but it also highlights possible behavioral interventions to reduce risk. Additionally, using a chemical speciation and factor analysis approach, sources of personal PM2.5, such as fresh sea salt, aged sea salt, and to a lesser degree, SHS and crustal sources, were associated with risk of decreased birthweight. Furthermore, high-loading components of these personal PM2.5 sources or factors and exposures highly correlated with them, may drive these associations, with magnesium and sodium most inversely associated with birthweight. Finally, I provide evidence that SHS, a major contributing source of personal PM2.5, may not be accurately measured by self-reported questionnaires which are typically used in pregnancy studies, supporting a call for standardized SHS exposure questionnaires to increase accuracy in exposure assessment and allow for harmonization and pooling of data and collaboration across cohorts. Overall, this dissertation highlights a need for more health research and targeted regulation and/or guidelines using PM2.5 sources and components considering PM2.5 is a complex mixture of organic and inorganic particles, which seem to differ in toxicity.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Personal PM2.5 exposure during pregnancy in an environmental health disparities population

PDF

Prenatal environmental exposures and fetal growth in the MADRES cohort

PDF

Assessment of the mortality burden associated with ambient air pollution in rural and urban areas of India

PDF

Urban air pollution and environmental justice: three essays

PDF

Sensor-based mobile health approaches for personal air pollution and pediatric asthma studies

PDF

Associations between ambient air pollution and hypertensive disorders of pregnancy

PDF

Air pollution, mitochondrial function, and growth in children

PDF

Examining exposure to extreme heat and air pollution and its effects on all-cause, cardiovascular, and respiratory mortality in California: effect modification by the social deprivation index

PDF

Assessing the impact of air pollution on adverse birth outcomes in a low resource setting

PDF

Spatial analysis of PM₂.₅ air pollution in association with hospital admissions in California

PDF

Effectiveness of individual- and household-level protective actions in reducing symptoms associated with hydrogen sulfide chronic low-level exposure

PDF

Air pollution, smoking, and multigenerational DNA methylation Signatures: a study of two southern California cohorts

$Exposure assessment and source apportionment of size fractions of airborne particulate matter$

PDF

Exposure assessment and source apportionment of size fractions of airborne particulate matter

PDF

The interplay between tobacco exposure and polygenic risk score for growth on birthweight and childhood acute lymphoblastic leukemia

PDF

Early life air pollution exposure, gestational diabetes mellitus, and autism spectrum disorder

PDF

Evaluation of new methods for estimating exposure to traffic-related pollution and early health effects for large population epidemiological studies

PDF

Spatial modeling of non-tailpipe emissions and its association with children's lung function

PDF

The role of pesticide exposure in breast cancer

PDF

Chemical and toxicological characteristics of particulate matter in urban environments with a focus on its sources, associated health impacts and mitigation policies

PDF

Prenatal exposure to organophosphate esters and early neurodevelopment

Asset Metadata

Creator O'Sharkey, Karl (author)

Core Title Personal exposure to particulate matter PM2.5 sources during pregnancy and birthweight

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Epidemiology

Degree Conferral Date 2022-08

Publication Date 07/22/2022

Defense Date 07/22/2022

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Air pollution,birthweight,Environmental Health,Epidemiology,measurement error,OAI-PMH Harvest,personal monitoring,PM2.5,Pregnancy,prenatal exposure

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Habre, Rima (committee chair), Breton, Carrie (committee member), Cockburn, Myles (committee member), Franklin, Meredith (committee member), Toledo-Corral, Claudia (committee member)

Creator Email karlosharkey@gmail.com,osharkey@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-oUC111373817

Unique identifier UC111373817

Legacy Identifier etd-OSharkeyKa-10895

Document Type Dissertation

Format application/pdf (imt)

Rights O'Sharkey, Karl

Type texts

Source 20220722-usctheses-batch-959 (batch), University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the author, as the original true and official version of the work, but does not grant the reader permission to use the work if the desired use is covered by copyright. It is the author, as rights holder, who must provide use permission if such use is covered by copyright. The original signature page accompanying the original submission of the work to the USC Libraries is retained by the USC Libraries and a copy of it may be obtained by authorized requesters contacting the repository e-mail address given.

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA

Repository Email cisadmin@lib.usc.edu