University of Southern California Dissertations and Theses

The effects of tobacco exposure on hormone levels and breast cancer risk among young women

(USC Thesis Other)

The effects of tobacco exposure on hormone levels and breast cancer risk among young women

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

THE EFFECTS OF TOBACCO EXPOSURE ON HORMONE LEVELS AND BREAST
CANCER RISK AMONG YOUNG WOMEN

By

Ugonna Nnebuaku Ihenacho

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

May 2021

Copyright 2021 Ugonna Nnebuaku Ihenacho
ii
Acknowledgements
First, I thank God for the strength, wisdom and love He gives me which allows me to do
all that I do. I am so blessed to have amazingly supportive family and friends by my side while
completing this research. Thank you to my family, Mom, Dad, Chukwuemeka, Kelechi, James,
Ifeanyi, Amauche, Chinemerem and Eche for your constant love and encouragement. Thank you
Uzo, Ikechi, Chimzorom and Maduka for being my little bundles of joy and stress relief.
I deeply appreciate my cohort, Ashley, Intira, Zhi, Malcolm, Zhaohui and Hua for the
screening exam study days, being a sounding board, the big laughs, and tasty dinners. I gratefully
acknowledge the late-night study warriors who would be with me in Soto and go on midnight
Starbucks runs so that I could be productive but safe. Carol, Sabrina, Karen, Kathy, and
Charlotte: I could not have done this without you! Thank you to my high school, Golden Bear
and MPH friends: you gave me so much love and inspiration when I needed it most. Thank you
to the numerous other friends and family that are not named here but inscribed on my heart.
Thank you to my amazing dissertation committee. Special thanks to my chair, Dr. Ann
Hamilton, and my co-chair, Dr. Anna Wu, for your guidance and for challenging me along the
way. Thank you to Dr. Wendy Mack for your advice on my analysis and thoughtful comments
on my drafts. Thank you, Dr. Jennifer Unger and Dr. Michael Press, for your time and
contributions, which helped me foster a better understanding of some very new topics.
Thank you to Dr. Ellen Velie and the YWHHS team for allowing me to work with your
data and having such valuable insight at each of our meetings. Thank you to my Preventive
Medicine instructors and advisors, especially Dr. Victoria Cortessis and Dr. Roberta McKean-
Cowdin.
And thank you, Jorge, for all things big and small. 143.
iii
Table of Contents
Acknowledgements ....................................................................................................................................... ii
List of Tables ............................................................................................................................................... vi
List of Figures ............................................................................................................................................ viii
Abstract ........................................................................................................................................................ ix
CHAPTER ONE: Background: Breast Cancer Incidence, Mortality, and Risk Factors ............................... 1
1.1 Incidence & mortality ............................................................................................................... 1
1.2 Genetic risk factors ................................................................................................................... 3
1.3 Sex steroid hormones ................................................................................................................ 5
1.4 Breast cancer subtypes .............................................................................................................. 5
1.5 Lifestyle and environmental risk factors ................................................................................... 7
1.6 Tobacco exposure and breast cancer risk ................................................................................ 10
1.7 Summary ................................................................................................................................. 13
1.8 References ............................................................................................................................... 14
CHAPTER TWO: A Systematic Review and Meta-Analysis of Smoking and Circulating Sex-Hormone
Levels Among Premenopausal Women ...................................................................................................... 29
2.1 Abstract ................................................................................................................................... 29
2.2 Introduction ............................................................................................................................. 30
2.3 Methods ................................................................................................................................... 31
2.4 Results ..................................................................................................................................... 35
2.5 Discussion ............................................................................................................................... 39
2.6 References ............................................................................................................................... 44
CHAPTER THREE: Using the Young Women’s Health History Study to Investigate the Role of Tobacco
Exposure and Breast Cancer Risk Among Young Women ........................................................................ 63
3.1 Study overview ....................................................................................................................... 63
3.2 Case / control definition and eligibility criteria ...................................................................... 63
iv
3.3 Data collection ........................................................................................................................ 64
3.4 Selection of YWHHS for dissertation research ...................................................................... 65
3.5 Dissertation chapters based on YWHHS: An overview ......................................................... 66
3.6 Variable definitions ................................................................................................................. 67
3.7 Summary ................................................................................................................................. 72
3.8 References ............................................................................................................................... 74
CHAPTER FOUR: Lifetime cigarette smoking and risk for breast cancer subtypes among Non-Hispanic
Black and White women in the Young Women’s Health History Study .................................................... 76
4.1 Abstract ................................................................................................................................... 77
4.2 Introduction ............................................................................................................................. 77
4.3 Methods ................................................................................................................................... 79
4.4 Results ..................................................................................................................................... 85
4.5 Discussion ............................................................................................................................... 87
4.6 References ............................................................................................................................... 92
CHAPTER FIVE: Lifetime composite smoking and risk for breast cancer subtypes among Non-Hispanic
Black and White women in the Young Women’s Health History Study .................................................. 112
5.1 Abstract ................................................................................................................................. 113
5.2 Introduction ........................................................................................................................... 113
5.3 Methods ................................................................................................................................. 115
5.4 Results ................................................................................................................................... 121
5.5 Discussion ............................................................................................................................. 124
5.6 References ............................................................................................................................. 127
CHAPTER SIX: Summary and Future Research ..................................................................................... 139
6.1 Summary and Conclusion ..................................................................................................... 139
6.2 Implications and Future Directions ....................................................................................... 140
6.3 References ............................................................................................................................. 143
v
APPENDIX ............................................................................................................................................... 144
Appendix 1. YWHHS questionnaire items used to derive tobacco exposure variables ............... 145

vi
List of Tables
Table 1.1 Summary of breast cancer susceptibility genes of high and moderate penetrance
Table 1.2 Summary of risk factors by breast cancer subtype
Table 2.1 Characteristics of studies selected for inclusion with measures of hormone levels
by smoking status
Table 2.2 Summary of the standardized mean differences between smokers and nonsmokers
for random effects among studies selected for meta-analysis by hormone
S. Table 2.1
1
PubMed search terms and results
S. Table 2.2 Equations for conversion from Wan et.al., 2014
S. Table 2.3 Summary of the standardized mean differences between smokers and nonsmokers
for random effects among studies with covariate-adjusted measures selected for
meta-analysis by hormone
Table 4.1 Characteristics of the control participants in the Young Women’s Health History
Study by personal smoking status
Table 4.2 Smoking characteristics of women in the Young Women’s Health History Study
by breast cancer status and subtype
Table 4.3 Multivariable-adjusted odds ratios and 95% CI for the association of personal
smoking history and risk of breast cancer overall and by tumor subtype
Table 4.4 Multivariable-adjusted odds ratios and 95% CI for the association of personal
smoking history and risk of breast cancer by race and HHP

1
S. Table is an abbreviation for Supplementary Table.
vii
S. Table 4.1 Association of personal smoking history and risk of breast cancer by BC subtype
and race
S. Table 4.2 Association of personal smoking history and risk of breast cancer by BC subtype
and HHP
S. Table 4.3 Sensitivity analyses estimating the association between ever smoking and BC risk
in listed subpopulations
Table 5.1 Sample weighted characteristics of the 1,381 controls in the Young Women’s
Health History Study by prenatal and passive smoke exposure status
Table 5.2 Lifetime smoking characteristics of the 3,193 women in the Young Women’s
Health History Study by breast cancer status and subtype
Table 5.3 Association lifetime smoking exposures and breast cancer risk, overall and by
tumor subtype
Table 5.4 Association between lifetime smoking history and breast cancer risk among
women without prenatal smoke exposure
Table 5.5 Association between lifetime smoking history and breast cancer risk by race and
SEP

viii
List of Figures
Figure 1.1 Summary of the pathways of estrogen-related carcinogenesis
Figure 1.2 Breast cancer incidence by subtype among NH White and NH Black women
under 50 years of age, SEER 21 2010-2016
Figure 2.1 Flowchart of selection process for articles included in assessment of smoking and
sex hormone level
Figure 2.2 Forest plots depicting the standardized mean differences between smokers and
nonsmokers for random effects among studies selected for meta-analysis by
hormone
Figure 2.3 Funnel plots depicting bias in the standardized mean differences by menstrual
phase and hormone
Figure 2.4 Power curves for estradiol by degree of heterogeneity and menstrual phase
S. Figure 2.1 Forest plots depicting the standardized mean differences between smokers and
nonsmokers for random effects among studies with covariate-adjusted measures
selected for meta-analysis by hormone

ix
Abstract
Breast cancer is the most common form of cancer diagnosed among women ages 20 to 49
years in the United States. While multiple risk factors for breast cancer have been identified, the
evidence of the role of tobacco exposure in breast cancer etiology has been conflicting. Recent
studies among premenopausal women have suggested that smoking is associated with an
increased risk of breast cancer. Still, additional studies are required to confirm this association
among populations of young women (under age 50 years), with a description of findings by
breast cancer subtype, and to evaluate differences in the association by race and socioeconomic
position (SEP). This dissertation aims to add to the growing literature evaluating the association
between tobacco exposure and breast cancer risk among young women, and on the potential
effect that tobacco exposure may have on endogenous hormone levels.
Background information on breast cancer incidence, mortality and risk factors are
provided in Chapter 1. Notable differences in breast cancer epidemiology are evident when
stratified by age (i.e., women age under vs over 50 years), by race, and by SEP. These
observations suggest differences in risk factors between women of a younger versus older age at
breast cancer diagnosis and provide context for the results of the analyses conducted in this
dissertation.
Three papers are included in this document: 1) a systematic review and meta-analysis of
endogenous hormone levels by smoking status among premenopausal women; 2) an evaluation
of the association between personal smoking history and breast cancer risk among young non-
Hispanic (NH) Black and NH White women using data from the Young Women’s Health
History Study (YWHHS), including an assessment of the association by race and SEP; and 3) an
evaluation of the association between lifetime composite smoke exposure (including prenatal,
x
personal, and secondhand exposure) and breast cancer risk, also using data from the YWHHS
with an evaluation of the association by race and SEP.
Chapter 2 presents a meta-analysis comprised of 19 studies evaluating the relationship
between sex-steroid hormone levels and smoking status. Specifically, estradiol, progesterone,
testosterone, dehydroepiandrosterone (DHEA), dehydroepiandrosterone sulfate (DHEAS), and
sex hormone binding globulin (SHBG) levels are described, as these hormones and proteins can
also be measures of estrogen bioavailability and metabolism. Because of the natural fluctuation
in hormone levels over a woman’s menstrual cycle, the findings of the meta-analysis are
presented by menstrual phase (follicular, luteal, varied) and overall. This paper also discusses the
hormonal mechanisms by which smoking may be associated with breast cancer risk.
In Chapter 3, the methodology of the Young Women’s Health History Study (YWHHS)
is described. This population-based case-control study collected information to evaluate risk
factors associated with invasive breast cancer among young NH Black and NH White women
under age 50 years residing in the Los Angeles County and Metropolitan Detroit Surveillance,
Epidemiology, and End Results (SEER) Registry areas. Controls were frequency-matched to
cases by region, five-year age group and race. The study utilized a complex sampling design to
allow for an analysis with sampling weights for the cases and controls that reflect the probability
of being selected into the study with an adjustment for nonresponse. The data from this uniquely
designed study are used for the two analyses described in Chapters 4 and 5.
Using data from the YWHHS, Chapter 4 discusses analyses of personal smoking history
and breast cancer risk in young women. Factors in smoking initiation, frequency, and intensity in
association with breast cancer risk are evaluated using a sample-weighted multivariable logistic
regression. The associations are explored by breast cancer status (control or case) and by
xi
molecular subtype: (luminal A, luminal B, human epidermal growth factor receptor 2-type, and
triple negative breast cancer). We also estimate the association stratified by race and SEP to
assess how differences in tobacco exposure patterns may contribute to the noted disparities in
breast cancer risk between young NH Black and NH White women.
Chapter 5 examines lifetime secondhand and lifetime composite smoke exposures in
relation to breast cancer risk. We explore differences in the association by prenatal smoke
exposure status, therefore considering the effects of smoking exposures across the participant’s
life span.
This analysis may provide insights into the biological mechanisms by which tobacco
exposure may impact breast cancer risk. Using the unique YWHHS dataset, we describe the
associations of lifetime tobacco exposures with breast cancer risk in subpopulations of NH Black
and NH White women and women of low SEP and mid/high SEP. By examining whether
tobacco exposure is associated with breast cancer risk according to breast cancer subtype, it may
be possible to generate knowledge that will motivate further investigation into pathways of
smoking carcinogenesis and justify targeted health education and tobacco-cessation interventions
to reduce breast cancer incidence among young women.

1
CHAPTER ONE: Background: Breast Cancer Incidence, Mortality, and Risk Factors

Breast cancer is the cancer most commonly diagnosed among women worldwide, overall
and among women under and over 50 years of age (1). The etiology and prognosis of a breast
cancer diagnosis vary by age, race, socioeconomic status, and breast cancer subtype. To inform
efforts in reducing breast cancer incidence in the diverse American population and populations
worldwide, it is important to understand the factors contributing to differences in breast cancer
risk among young women (i.e., women diagnosed under 50 years of age) versus older women,
and how these differences vary by race and socioeconomic status.

1.1 Incidence & mortality
Breast cancer is the most common cancer among women in the United States (US) with
an estimate of 268,600 new cases in 2019 (1). The age-standardized incidence rate for female
breast cancer in the US is 127.5 per 100,000 person-years based on Surveillance, Epidemiology,
and End Results (SEER) Program data from 2012-2016 (1). Also, in that time the breast cancer
incidence rose an average of 0.5% each year for both Black and White women (1).
Approximately one in eight women will be diagnosed with breast cancer in their lifetime (1).
Although breast cancer is the most commonly occurring cancer site, breast cancer is the
3
rd
most common cause of cancer-related death among women, behind lung and colorectal
cancer (1). SEER estimates that 20.6 women per 100,000 person-years will die of breast cancer,
with roughly 41,760 deaths expected in 2019 (1). Long-term efforts in early screening and
treatment advances have contributed to an average annual decrease in breast cancer mortality of
1.5% among non-Hispanic White women and a 1% average annual decrease among non-
2
Hispanic Black women from 2013-2017 (1; 2). Approximately 3.5 million women in the US in
2016 were living with breast cancer (1).

1.1.1 Among young women (ages 20-49 years)
There are some notable differences in breast cancer incidence and mortality among
women under age 50 years compared to women aged 50 years and older. Since breast cancer
incidence increases with age, the overall rate is lower among those <50 than among women ≥50
years. However, breast cancer has the highest incidence and mortality rates of all cancers
diagnosed among young women ages 20-49 years (3). According to SEER data from 2011-2016,
breast cancer incidence and mortality among young women was 73.2 per 100,000 person-years
and 7.3 per 100,000 person-years, respectively. The incidence and mortality rates of breast
cancer in young women are more than twice the respective rates for any other cancer diagnosed
among young women (3). Compared to the trends in women of all ages, younger women
experienced a slightly higher increase in incidence (0.5% 5-year average annual percent change
for 2001-2015 vs. 0.3% over all ages) and similar mortality rate (-1.3% 5-year average annual
percent change for those < 50 years and overall) between 1999-2016 (1; 3).

1.1.2 Racial differences among non-Hispanic White and non-Hispanic Black younger
women
Breast cancer is the most commonly diagnosed cancer and the second most deadly cancer
diagnosed among Black women of any age (4; 5). Although the overall age-adjusted breast
cancer incidence is lower among Black women compared to non-Hispanic (NH) White women,
Black women 20-40 years of age have a breast cancer incidence that is 17% higher (incidence:
3
38.8 vs. 27.9 per 100,000 persons in 2000-2017) and a breast cancer mortality twice that of NH
White women of the same age group (mortality: 4.2 vs. 9.4 per 100,000 deaths in 2000-2017) (1;
5-7). Approximately one in eight Black women newly diagnosed with breast cancer is under age
40, whereas the estimate is about one in twenty among NH White women (4; 6).

1.1.3 Breast cancer and socioeconomic position
In a study conducted using data from the Los Angeles County Cancer Surveillance
Program, socioeconomic factors, as developed from census tract data, were positively associated
with breast cancer risk across five racial/ethnic groups (NH White, Black, Hispanic White, Asian
and Other) and across ten-year age groups (8). Overall, breast cancer risk increased with
increasing socioeconomic status (SES; based on census data) (Ptrend=0.0001); women in the
highest quintile of SES had a 53% increased risk compared to women in the lowest quintile
(relative risk (RR) 1.53; 95% confidence interval (CI): 1.49, 1.57) (8). Multiple studies around
the world have observed similar results (9-12).
Low socioeconomic position (SEP; as measured by social and economic factors like
insurance status and poverty) is also associated with increased mortality (10; 13; 14). Low breast
cancer screening adherence, treatment delays and insufficient treatment have been identified as
factors contributing to disparities in mortality by SEP; however, even after adjusting for
healthcare factors and cancer characteristics at diagnosis, the association remains (15; 16).

1.2 Genetic risk factors
Genetic factors are estimated to contribute to 5-10% of all breast cancers (17-19). Recent
studies suggest that genetic mutations of high and moderate risk may be present in 12-23% of
4
breast cancer cases diagnosed among women under age 50 (20; 21). Mutations in the BRCA1 and
BRCA2 genes confer a high risk for heritable breast cancer (22; 23). The BRCA1/2 genes are
tumor-suppressor genes that code for proteins involved in deoxyribonucleic acid (DNA) repair.
In a meta-analysis of 10 studies, the cumulative risk of breast cancer by age 70 years was 57%
(95% CI: 47%, 66%) among women carrying mutations in BRCA1 and 49% (95% CI: 40%,
57%) for BRCA2 mutation carriers (23). It is estimated that about 5% of breast cancer cases carry
mutations in the BRCA1/2 gene, with a prevalence of about 4-10% among White women
diagnosed under age 40 years (24). A prevalence of 25-67% has been identified among
Ashkenazi Jewish women and 17% among Black women with a breast cancer diagnosis at age
35 years or younger (24-26). More than 3,000 pathogenic or likely pathogenic variants in BRCA1
and 3,400 in BRCA2 have been identified, with ongoing studies to identify additional mutations
in various racial/ethnic subgroups (27-34).
Studies of the DNA repair pathway and function of the BRCA1/2 genes and proteins have
contributed to the discovery of other breast cancer susceptibility genes of high and moderate
penetrance (probability of breast cancer susceptibility gene carriers to be diagnosed with breast
cancer; Table 1.1) (22; 35; 36). Carriers of these breast cancer susceptibility genes have a risk of
breast cancer that is 2 to 6 times that of non-carriers (19). Carrying mutations in TP53 can confer
a risk more than 100 times that of non-carriers (RR: 105; 90% CI: 62, 165) (19). Other
susceptibility genes and genes of low penetrance include: RAD50, CDKN2A, MSH2, MLH1,
MSH6, PMS2, MUTYH, BARD1, FAM175A, MRE11A, RAD51C (22; 35; 36). Furthermore,
mutations in TOX3, ESR1, FGFR2, RAD51, CCND1, SLC4A7 and MAP3K1 have been
suggested to be associated with young-onset breast cancer (37-42).

5
1.3 Sex steroid hormones
Endogenous levels of estrogens, testosterone, androstenedione, and other androgens have
been identified as hormonal risk factors for breast cancer, especially via estrogenic pathways
(43-47). Testosterone, androstenedione and other androgens can be converted to estrone (E1) and
estradiol (E2) by aromatase to increase endogenous estrogen levels (43).
Through metabolic pathways described by Yager and Davidson (Figure 1.1), the parent
estrogens (E1 or E2) are oxidized to form hydroxylated metabolites which then contribute to
estrogen-related carcinogenesis by: 1) binding to proteins and DNA; and 2) creating quinones
that form adducts that lead to DNA depurination or form reactive oxygen species that contribute
to oxidative DNA damage (43; 48). Activation of the estrogen receptor protein is also implicated
in cell proliferation and inhibited apoptosis, as this process has the potential to activate signaling
pathways that lead to altered gene expression (43).
Studies of exogenous hormone use among postmenopausal women have found that even
five years of combined estrogen plus progestin postmenopausal hormone therapy use was
associated with approximately a 25% increased risk of breast cancer (49-51). A recent meta-
analysis estimated an 8% increased risk of breast cancer with a history of estrogen plus progestin
oral contraceptive use (51; 52). Current and long-term combined oral contraceptive use (≥10
years of use compared to no use) was also associated with a 30-70% increased risk of breast
cancer among young women (53-56).

1.4 Breast cancer subtypes
Breast cancers are characterized by the tumor expression of two hormone receptors (HR),
the estrogen receptor (ER) and progesterone receptor (PR), the human epidermal growth factor
6
receptor 2 (HER2), and grade of the tumor. The four major molecular subtypes are described as
(57):
• Luminal A: ER+ and/or PR+, HER2-, low grade;
• Luminal B: ER+ and/or PR+, HER2-/+, higher grade;
• HER2-enriched: ER-, PR-, HER2+ (any grade);
• Triple negative breast cancer (TNBC): ER-, PR-, HER2- (any grade).
Characterization of the Ki67 protein, a marker of cell proliferation, is sometimes used to
distinguish between luminal subtypes (57). An index of less than 13.25% (clinical cut point of
14%) designates luminal A tumors and higher values designate luminal B tumors (57).
TNBCs are a heterogeneous group of breast cancers that can be further classified as
basal-like 1, basal-like 2, mesenchymal and luminal androgen receptor TNBC subtypes (58).
Although the terms basal-like and TNBC are often used interchangeably, not all TNBCs are
basal like; basal-like TNBCs tend to be high grade invasive ductal carcinomas (57). Women
diagnosed with TNBC are more likely to be premenopausal than postmenopausal breast cancer
cases (9).
The characterization of the breast cancer molecular subtype and histology are used to
inform treatment options and patient prognosis. Luminal A breast cancer subtypes make up about
67% of all female breast cancers diagnosed in the US and have the most favorable prognosis of
the subtypes with a 94% 5-year survival rate; Luminal B breast cancers account for 10% of all
breast cancers and about 89.8% of women diagnosed with this subtype will survive to 5 years
post-diagnosis; HER2-enriched cancers constitute 4% and have an 83% survival rate; and TNBC
constitute 10% and are associated with the poorest prognosis with a 76.5% 5-year survival rate
(59). Beyond these 4 major molecular categories, breast cancers can be classified as one of at
7
least 20 histologic breast cancer subtypes characterized by their morphology, clinical behaviors,
and molecular features (57; 59).

1.4.1 Breast cancer subtype by race and SEP
Breast cancer subtype varies by race (Figure 1.2) (7). Premenopausal Black women
diagnosed with basal-like TNBC tumors were two times more likely to develop rapidly growing
and proliferating disease than White women diagnosed with basal-like TNBC tumors (6; 60).
Additionally, individual SEP (based on increasing education level) and neighborhood
SEP (based on quintiles of neighborhood SES) were positively associated with ER+ breast
cancer risk in a 2012 study from the Black Women’s Health Study (BWHS; Ptrend =0.002; Ptrend
=0.02, respectively) (61). However, the observed trends weakened after adjusting for
reproductive factors and the study did not differentiate between post- and pre-menopausal breast
cancer (61). In a study conducted by the California Cancer Registry, women of mid- to low-SEP
were 10-20% more likely to be diagnosed with TNBC than other breast cancer subtypes (9).

1.5 Lifestyle and environmental risk factors
Many studies, including the California Teachers Study, the Multiethnic Cohort Study,
and the Nurses’ Health Study, have identified and described breast cancer risk factors. The
strongest risk factors that are not modifiable include older age and being born female (62). In the
US, the median age at breast cancer diagnosis among women is 61 years (63). Factors in a
woman’s medical history that are positively associated with breast cancer risk include a family
history of breast cancer, history of benign breast disease, and higher breast density (62). Older
age at first full-term pregnancy, nulliparity, decreased duration of breastfeeding, early age of
8
menarche, and older age at menopause are some menstrual/reproductive factors associated with
an increased risk of breast cancer (62). On the other hand, young age at first full-term pregnancy,
an increasing number of full-term pregnancies, and long duration of breastfeeding are associated
with a decreased risk of breast cancer (63).
Risk factors that are considered modifiable and positively associated with breast cancer
risk include obesity, weight gain, oral contraceptive use, physical inactivity, and alcohol
consumption (62). In a study of postmenopausal women, those with a body mass index (BMI)
greater than 35 kg/m
2
were 1.58 times as likely to develop breast cancer compared to women of
normal BMI of 20-24.9 kg/m
2
(95% CI: 1.40, 1.79) (64). Conversely, the California Teacher’s
Study identified that consumption of a plant-based diet in the highest quintile was associated
with a 15% decreased risk of breast cancer compared to women in the lowest quintile (RR: 0.85:
95% CI: 0.76, 0.95; Ptrend = 0.003) (65).
Among premenopausal populations, reproductive factors that have been associated with
increased breast cancer risk include early age at menarche and first birth at age 30 years or more;
older age at menarche and breastfeeding among parous women confer decreased risk (66-68).
Contrary to the association observed among postmenopausal women, obesity is associated with a
decreased risk of breast cancer in premenopausal populations with one study identifying a 25%
decreased risk of premenopausal breast cancer among overweight and obese women with a BMI
≥25 kg/m
2
compared to women with a normal BMI (69; 70). It has been suggested that high
levels of physical activity (compared to lowest level) confer greater protection from breast
cancer among premenopausal women (RR: 0.77; 95% CI: 0.72, 0.84) than postmenopausal
women (RR: 0.88; 95% CI: 0.84, 0.92) (71-73).

9
1.5.1 Socioeconomic position, race, and protective factors
A recent study suggests that SEP and reproductive factors (parity and age at first birth)
explain much of the racial disparity in the risk of hormone receptor (ER/PR) negative disease
(74). SEP is positively associated with hormone receptor positive breast cancer (61). Older age at
first full-term pregnancy and low parity are commonly observed among women with high SEP
and both are risk factors for hormone receptor positive breast cancer (8; 75; 76).
A combination of parity and breastfeeding may contribute to racial and socioeconomic
differences in breast cancer risk. A study of TNBC identified that young parous Black women
that breastfed for six months or more had an 82% decreased risk of breast cancer compared to
those that never breastfed (OR: 0.18; 95% CI: 0.07, 0.46) (77). Shorter duration of breastfeeding
is also associated with lower SEP, according to a 2015 study on workplace barriers and
breastfeeding (78). Other personal barriers to breastfeeding included decreased or no paid
maternity leave and a lack of workplace protections (78). Black women have the second highest
rate of being among the working poor, working at least 27 weeks of the year but being below the
poverty level (79). Black women are also more likely to be a single head of household compared
to women of other races, and more than 30% of households with a single female at the head had
incomes below the poverty level in 2012 (79). Considering that Black women are more likely to
be the single head of household and among the working poor, managing the financial well-being
of her family may contribute to racial disparities in breast cancer risk by precluding the
protective effects of a longer duration of breastfeeding.
The relationship between body size and breast cancer risk varies across the life course.
Multiple studies have described an inverse association between adolescent and young adult BMI
and premenopausal breast cancer risk (69; 80; 81). The findings among Black women have been
10
inconsistent but the African American Breast Cancer Epidemiology and Risk (AMBER)
Consortium identified that a high BMI in young adulthood was associated with a decreased risk
of premenopausal ER+ breast cancer among Black women (82). In a pooled analysis of 20
prospective studies, adult weight gain and adult BMI were associated with an increased risk of
postmenopausal breast and especially for HR+ breast cancer (81). These findings were also
confirmed among Black postmenopausal women in the AMBER Consortium (82).

1.5.2 Risk factors and breast cancer subtype among young women
Among women under age 50 years, an increasing number of full-term pregnancies is
inversely associated with the risk of ER+/PR+ breast cancer, average alcohol consumption in the
recent 5 years is associated with an increased risk of ER+/PR+ breast cancer, and being
overweight and obese is associated with a decreased risk of luminal breast cancer (83; 84).
Numerous studies have been conducted to identify the genetic, environmental, and
hormonal risk factors by breast cancer subtype (summarized in Table 1.2) (85; 86).

1.6 Tobacco exposure and breast cancer risk
There is accumulating evidence that personal smoking is associated with an increased
risk of breast cancer although results are not all consistent. In 2013, Gaudet et.al. published
results from the American Cancer Society’s Cancer Prevention Study II (ACS CPSII) study
concluding that breast cancer incidence was higher among current (HR: 1.24, 95% CI: 1.07 to
1.42) and former smokers (HR: 1.13, 95% CI: 1.06 to 1.21) compared to never smokers (87). A
meta-analysis of 30 studies in the 2014 Surgeon General’s Report described a 9% increased risk
of breast cancer among ever smokers (RR: 1.09; 95% CI: 1.06, 1.12; Pheterogeneity =0.50) (88). In
11
the Nurses’ Health Study, smoking was associated with an increased risk of premenopausal
breast cancer; HR was 1.11 (95% CI: 1.07-1.15) for each additional 20 pack-years of use (89). In
most of these studies, more than 80% of the participants were White and the Nurses’ Health
Study consisted of mostly professional women (90; 91).
In the BWHS, a cohort of 52,425 Black women, the association between tobacco
exposure and breast cancer risk was evaluated by lifetime active and secondhand smoking status
(92). Risk of premenopausal breast cancer was increased among smokers; the incidence rate ratio
(IRR) was 1.70 (95% CI 1.05–2.75) associated with initiating smoking before age 18 and having
at least 20 pack-years of smoking history compared to never active or passive smokers (92).
Based on these results, Rosenberg and colleagues suggested that Black women who smoke may
be at increased risk of breast cancer compared to White women who smoke, but this difference
warrants further evaluation.
The BWHS also found that secondhand smoke exposure was associated with an increased
risk of premenopausal breast cancer among Black women (IRR: 1.42; 95% CI 1.09–1.85) (92).
The 2014 Surgeon General’s Report meta-analysis, secondhand smoke exposure was associated
with a 21% increased risk of premenopausal breast cancer (RR: 1.21; 95% CI: 1.04, 1.40; N=12
studies; Pheterogeneity=0.063) (results from BWHS was not included in this meta-analysis) (88).
Recent studies have evaluated the timing of passive tobacco exposures in relation to breast
cancer risk. The Sister Study, a cohort study of women without breast cancer who had a sister
with breast cancer, identified that secondhand exposure before age 18 was associated with a 17%
increased risk of breast cancer (HR:1.17; 95% CI: 1.00, 1.36) compared to women with no
childhood secondhand smoke exposure (93). They also found that in utero exposure to household
secondhand smoke during a mother’s pregnancy was associated with a 16% increased risk of
12
breast cancer among their offspring (HR:1.16; 95% CI: 1.01, 1.32) (93). The study of the timing
of passive tobacco exposures is a growing area of breast cancer research, as inconsistent results
have been published and few of these studies have evaluated the association specifically for
young-onset breast cancer (94-101).

1.6.1 Tobacco exposure and breast cancer subtype
Current smoking was associated with a 2-fold increased risk of ER- breast cancer (RR:
2.21; 95% CI: 1.23, 3.96) in a study of pre- and postmenopausal women (102). Among a
population of women ages 20-44 years, smoking 10 pack-years or more was associated with a
60% increased risk of ER+ breast cancer (OR: 1.6; 95% CI: 1.1, 2.4) (103). These findings
suggest that there may be differences in breast cancer risk by menopausal status and breast
cancer subtype. More research on the association between smoking and breast cancer are needed
to support the developing research in the biological pathway of tobacco carcinogenesis in breast
cancer.

2.6.2 Tobacco exposure, race, and SEP
Studies on the association between tobacco exposure and breast cancer by race and SEP
are also warranted due to differences in tobacco exposures between White and Black women,
and between low- and high-SEP women. The Centers for Disease Control and Prevention (CDC)
reports that Black smokers tend to initiate smoking at a later age and smoke fewer cigarettes per
day compared to White smokers (104). However, it is important to also consider that slower
nicotine metabolism and higher nicotine intake per cigarette have been described among Black
smokers; these factors have been suggested to confer increased risk (105-107). Black children
13
and adults are more likely to be exposed to secondhand smoke than people of any other race
(104). Similarly, individuals living below the poverty level in the United States (US) are more
likely to be exposed to secondhand smoke; however, smokers of low SEP tend to smoke more
heavily compared to those above the poverty level (104).

1.7 Summary
Breast cancer among women under age 50 years is associated with some differences in
the environmental risk factors, genetic characteristics and presentation of the disease compared
to women diagnosed at an older age. With a growing US population and an increasing trend in
breast cancer incidence among young women, it can be expected that over time the absolute
number of young women with a breast cancer diagnosis will increase and these women will
require long-term health care to manage the effects of cancer or cancer treatment on their
survival and survivorship. Women under age 45 years with breast cancer have an estimated
12.78 years of life lost, a doubling of the estimate of 4.99 years for older women (age 15-44
years compared to 45-64 years; p<0.00025) (108). For those reasons, special attention is required
to identify risk factors associated with breast cancer among women under age 50 to be used in
primary prevention strategies and secondary prevention strategies to identify populations that
may require early screening for breast cancer.

14
1.8 References
1. DeSantis, C. E., Ma, J., Gaudet, M. M., Newman, L. A., Miller, K. D., Goding Sauer, A., . .
. Siegel, R. L. (2019). Breast cancer statistics, 2019. CA Cancer J Clin, 69(6), 438-451.
doi:10.3322/caac.21583
2. Cronin, K. A., Lake, A. J., Scott, S., Sherman, R. L., Noone, A. M., Howlader, N., . . .
Jemal, A. (2018). Annual Report to the Nation on the Status of Cancer, part I: National
cancer statistics. Cancer, 124(13), 2785-2800. doi:10.1002/cncr.31551
3. Ward, E. M., Sherman, R. L., Henley, S. J., Jemal, A., Siegel, D. A., Feuer, E. J., . . .
Cronin, K. A. (2019). Annual Report to the Nation on the Status of Cancer, Featuring
Cancer in Men and Women Age 20-49 Years. J Natl Cancer Inst, 111(12), 1279-1297.
doi:10.1093/jnci/djz106
4. American Cancer Society. (2016). Cancer Facts & Figures for African Americans 2016-
2018. https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-
statistics/cancer-facts-and-figures-for-african-americans/cancer-facts-and-figures-for-
african-americans-2016-2018.pdf
5. Swanson, G. M., Haslam, S. Z., & Azzouz, F. (2003). Breast cancer among young African-
American women: a summary of data and literature and of issues discussed during the
Summit Meeting on Breast Cancer Among African American Women, Washington, DC,
September 8-10, 2000. Cancer, 97(1 Suppl), 273-279. doi:10.1002/cncr.11025
6. Danforth, D. N., Jr. (2013). Disparities in breast cancer outcomes between Caucasian and
African American women: a model for describing the relationship of biological and
nonbiological factors. Breast Cancer Res, 15(3), 208. doi:10.1186/bcr3429
7. National Cancer Institute Surveillance Research Program. National Cancer Institute
SEER*Stat software. Retrieved September 15, 2019 https://seer.cancer.gov/seerstat
version 8.3.6
8. Liu, L., Deapen, D., & Bernstein, L. (1998). Socioeconomic status and cancers of the
female breast and reproductive organs: a comparison across racial/ethnic populations in
Los Angeles County, California (United States). Cancer Causes Control, 9(4), 369-380.
9. Bauer, K. R., Brown, M., Cress, R. D., Parise, C. A., & Caggiano, V. (2007). Descriptive
analysis of estrogen receptor (ER)-negative, progesterone receptor (PR)-negative, and
HER2-negative invasive breast cancer, the so-called triple-negative phenotype: a
population-based study from the California cancer Registry. Cancer, 109(9), 1721-1728.
doi:10.1002/cncr.22618
10. Beiki, O., Hall, P., Ekbom, A., & Moradi, T. (2012). Breast cancer incidence and case
fatality among 4.7 million women in relation to social and ethnic background: a
population-based cohort study. Breast Cancer Res, 14(1), R5. doi:10.1186/bcr3086
15
11. Bhurgri, Y., Kayani, N., Faridi, N., Pervez, S., Usman, A., Bhurgri, H., . . . Zaidi, S. H.
(2007). Patho-epidemiology of breast cancer in Karachi '1995-1997'. Asian Pac J Cancer
Prev, 8(2), 215-220.
12. Borugian, M. J., Spinelli, J. J., Abanto, Z., Xu, C. L., & Wilkins, R. (2011). Breast cancer
incidence and neighbourhood income. Health Rep, 22(2), 7-13.
13. Tian, N., Goovaerts, P., Zhan, F. B., Chow, T. E., & Wilson, J. G. (2012). Identifying risk
factors for disparities in breast cancer mortality among African-American and Hispanic
women. Womens Health Issues, 22(3), e267-276. doi:10.1016/j.whi.2011.11.007
14. Okunade, A. A., & Karakus, M. C. (2003). Mortality from breast carcinoma among US
women: the role and implications of socio-economics, heterogeneous insurance, screening
mammography, and geography. Health Care Manag Sci, 6(4), 237-248.
15. Silber, J. H., Rosenbaum, P. R., Ross, R. N., Reiter, J. G., Niknam, B. A., Hill, A. S., . . .
Fox, K. R. (2018). Disparities in Breast Cancer Survival by Socioeconomic Status Despite
Medicare and Medicaid Insurance. Milbank Q, 96(4), 706-754. doi:10.1111/1468-
0009.12355
16. Dreyer, M. S., Nattinger, A. B., McGinley, E. L., & Pezzin, L. E. (2018). Socioeconomic
status and breast cancer treatment. Breast Cancer Res Treat, 167(1), 1-8.
doi:10.1007/s10549-017-4490-3
17. Martin, A. M., & Weber, B. L. (2000). Genetic and hormonal risk factors in breast cancer.
J Natl Cancer Inst, 92(14), 1126-1135. doi:10.1093/jnci/92.14.1126
18. Claus, E. B., Schildkraut, J. M., Thompson, W. D., & Risch, N. J. (1996). The genetic
attributable risk of breast and ovarian cancer. Cancer, 77(11), 2318-2324.
doi:10.1002/(SICI)1097-0142(19960601)77:11<2318::AID-CNCR21>3.0.CO;2-Z
19. Tung, N., Lin, N. U., Kidd, J., Allen, B. A., Singh, N., Wenstrup, R. J., . . . Garber, J. E.
(2016). Frequency of Germline Mutations in 25 Cancer Susceptibility Genes in a
Sequential Series of Patients With Breast Cancer. J Clin Oncol, 34(13), 1460-1468.
doi:10.1200/JCO.2015.65.0747
20. Rummel, S. K., Lovejoy, L., Shriver, C. D., & Ellsworth, R. E. (2017). Contribution of
germline mutations in cancer predisposition genes to tumor etiology in young women
diagnosed with invasive breast cancer. Breast Cancer Res Treat, 164(3), 593-601.
doi:10.1007/s10549-017-4291-8
21. Copson, E. R., Maishman, T. C., Tapper, W. J., Cutress, R. I., Greville-Heygate, S.,
Altman, D. G., . . . Eccles, D. M. (2018). Germline BRCA mutation and outcome in young-
onset breast cancer (POSH): a prospective cohort study. Lancet Oncol, 19(2), 169-180.
doi:10.1016/S1470-2045(17)30891-4
16
22. Shuen, A. Y., & Foulkes, W. D. (2011). Inherited mutations in breast cancer genes--risk
and response. J Mammary Gland Biol Neoplasia, 16(1), 3-15. doi:10.1007/s10911-011-
9213-5
23. Chen, S., & Parmigiani, G. (2007). Meta-analysis of BRCA1 and BRCA2 penetrance. J
Clin Oncol, 25(11), 1329-1333. doi:10.1200/JCO.2006.09.1066
24. Kurian, A. W. (2010). BRCA1 and BRCA2 mutations across race and ethnicity:
distribution and clinical implications. Curr Opin Obstet Gynecol, 22(1), 72-78.
doi:10.1097/GCO.0b013e328332dca3
25. John, E. M., Miron, A., Gong, G., Phipps, A. I., Felberg, A., Li, F. P., . . . Whittemore, A.
S. (2007). Prevalence of pathogenic BRCA1 mutation carriers in 5 US racial/ethnic groups.
JAMA, 298(24), 2869-2876. doi:10.1001/jama.298.24.2869
26. Robson, M., Dabney, M. K., Rosenthal, G., Ludwig, S., Seltzer, M. H., Gilewski, T., . . .
Offit, K. (1997). Prevalence of recurring BRCA mutations among Ashkenazi Jewish
women with breast cancer. Genet Test, 1(1), 47-51. doi:10.1089/gte.1997.1.47
27. Cherbal, F., Salhi, N., Bakour, R., Adane, S., Boualga, K., & Maillet, P. (2012). BRCA1
and BRCA2 unclassified variants and missense polymorphisms in Algerian breast/ovarian
cancer families. Dis Markers, 32(6), 343-353. doi:10.3233/DMA-2012-0893
28. Alhuqail, A. J., Alzahrani, A., Almubarak, H., Al-Qadheeb, S., Alghofaili, L., Almoghrabi,
N., . . . Karakas, B. (2018). High prevalence of deleterious BRCA1 and BRCA2 germline
mutations in arab breast and ovarian cancer patients. Breast Cancer Res Treat, 168(3), 695-
702. doi:10.1007/s10549-017-4635-4
29. Lopez-Urrutia, E., Salazar-Rojas, V., Brito-Elias, L., Coca-Gonzalez, M., Silva-Garcia, J.,
Sanchez-Marin, D., . . . Perez-Plasencia, C. (2019). BRCA mutations: is everything said?
Breast Cancer Res Treat, 173(1), 49-54. doi:10.1007/s10549-018-4986-5
30. Feng, Y., Rhie, S. K., Huo, D., Ruiz-Narvaez, E. A., Haddad, S. A., Ambrosone, C. B., . . .
Haiman, C. A. (2017). Characterizing Genetic Susceptibility to Breast Cancer in Women of
African Ancestry. Cancer Epidemiol Biomarkers Prev, 26(7), 1016-1026.
doi:10.1158/1055-9965.EPI-16-0567
31. Friebel, T. M., Andrulis, I. L., Balmana, J., Blanco, A. M., Couch, F. J., Daly, M. B., . . .
Rebbeck, T. R. (2019). BRCA1 and BRCA2 pathogenic sequence variants in women of
African origin or ancestry. Hum Mutat, 40(10), 1781-1796. doi:10.1002/humu.23804
32. Cortes, C., Rivera, A. L., Trochez, D., Solarte, M., Gomez, D., Cifuentes, L., & Barreto, G.
(2019). Mutational analysis of BRCA1 and BRCA2 genes in women with familial breast
cancer from different regions of Colombia. Hered Cancer Clin Pract, 17, 20.
doi:10.1186/s13053-019-0120-x
33. Landrum, M. J., Lee, J. M., Riley, G. R., Jang, W., Rubinstein, W. S., Church, D. M., &
Maglott, D. R. (2014). ClinVar: public archive of relationships among sequence variation
17
and human phenotype. Nucleic Acids Res, 42(Database issue), D980-985.
doi:10.1093/nar/gkt1113
34. Kim, Y. C., Zhao, L., Zhang, H., Huang, Y., Cui, J., Xiao, F., . . . Wang, S. M. (2016).
Prevalence and spectrum of BRCA germline variants in mainland Chinese familial breast
and ovarian cancer patients. Oncotarget, 7(8), 9600-9612. doi:10.18632/oncotarget.7144
35. Kurian, A. W., Hughes, E., Handorf, E. A., Gutin, A., Allen, B., Hartman, A.-R., & Hall,
M. J. (2017). Breast and Ovarian Cancer Penetrance Estimates Derived From Germline
Multiple-Gene Sequencing Results in Women. JCO Precision Oncology(1), 1-12.
doi:10.1200/po.16.00066
36. Kurian, A. W., Ward, K. C., Howlader, N., Deapen, D., Hamilton, A. S., Mariotto, A., . . .
Katz, S. J. (2019). Genetic Testing and Results in a Population-Based Cohort of Breast
Cancer Patients and Ovarian Cancer Patients. J Clin Oncol, 37(15), 1305-1315.
doi:10.1200/JCO.18.01854
37. Shi, M., O'Brien, K. M., Sandler, D. P., Taylor, J. A., Zaykin, D. V., & Weinberg, C. R.
(2017). Previous GWAS hits in relation to young-onset breast cancer. Breast Cancer Res
Treat, 161(2), 333-344. doi:10.1007/s10549-016-4053-z
38. Ahsan, H., Halpern, J., Kibriya, M. G., Pierce, B. L., Tong, L., Gamazon, E., . . .
Whittemore, A. S. (2014). A genome-wide association study of early-onset breast cancer
identifies PFKM as a novel breast cancer gene and supports a common genetic spectrum
for breast cancer at any age. Cancer Epidemiol Biomarkers Prev, 23(4), 658-669.
doi:10.1158/1055-9965.EPI-13-0340
39. Rath, M., Li, Q., Li, H., Lindstrom, S., Miron, A., Miron, P., . . . Ruddy, K. J. (2019).
Evaluation of significant genome-wide association studies risk - SNPs in young breast
cancer patients. PLoS One, 14(5), e0216997. doi:10.1371/journal.pone.0216997
40. Maxwell, K. N., Wubbenhorst, B., D'Andrea, K., Garman, B., Long, J. M., Powers, J., . . .
Nathanson, K. L. (2015). Prevalence of mutations in a panel of breast cancer susceptibility
genes in BRCA1/2-negative patients with early-onset breast cancer. Genet Med, 17(8),
630-638. doi:10.1038/gim.2014.176
41. Campeau, P. M., Foulkes, W. D., & Tischkowitz, M. D. (2008). Hereditary breast cancer:
new genetic developments, new therapeutic avenues. Hum Genet, 124(1), 31-42.
doi:10.1007/s00439-008-0529-1
42. Easton, D. F. (1999). How many more breast cancer predisposition genes are there? Breast
Cancer Res, 1(1), 14-17. doi:10.1186/bcr6
43. Yager, J. D., & Davidson, N. E. (2006). Estrogen carcinogenesis in breast cancer. N Engl J
Med, 354(3), 270-282. doi:10.1056/NEJMra050776
44. Spicer, D. V., & Pike, M. C. (1993). Breast cancer prevention through modulation of
endogenous hormones. Breast Cancer Res Treat, 28(2), 179-193. doi:10.1007/bf00666430
18
45. Kaaks, R., Berrino, F., Key, T., Rinaldi, S., Dossus, L., Biessy, C., . . . Riboli, E. (2005).
Serum sex steroids in premenopausal women and breast cancer risk within the European
Prospective Investigation into Cancer and Nutrition (EPIC). J Natl Cancer Inst, 97(10),
755-765. doi:10.1093/jnci/dji132
46. Eliassen, A. H., Missmer, S. A., Tworoger, S. S., Spiegelman, D., Barbieri, R. L., Dowsett,
M., & Hankinson, S. E. (2006). Endogenous steroid hormone concentrations and risk of
breast cancer among premenopausal women. J Natl Cancer Inst, 98(19), 1406-1415.
doi:10.1093/jnci/djj376
47. James, R. E., Lukanova, A., Dossus, L., Becker, S., Rinaldi, S., Tjonneland, A., . . . Kaaks,
R. (2011). Postmenopausal serum sex steroids and risk of hormone receptor-positive and -
negative breast cancer: a nested case-control study. Cancer Prev Res (Phila), 4(10), 1626-
1635. doi:10.1158/1940-6207.CAPR-11-0090
48. Yager, J. D. (2015). Mechanisms of estrogen carcinogenesis: The role of E2/E1-quinone
metabolites suggests new approaches to preventive intervention--A review. Steroids, 99(Pt
A), 56-60. doi:10.1016/j.steroids.2014.08.006
49. Rossouw, J. E., Anderson, G. L., Prentice, R. L., LaCroix, A. Z., Kooperberg, C.,
Stefanick, M. L., . . . Writing Group for the Women's Health Initiative, I. (2002). Risks and
benefits of estrogen plus progestin in healthy postmenopausal women: principal results
From the Women's Health Initiative randomized controlled trial. JAMA, 288(3), 321-333.
doi:10.1001/jama.288.3.321
50. Chlebowski, R. T., Hendrix, S. L., Langer, R. D., Stefanick, M. L., Gass, M., Lane, D., . . .
Investigators, W. H. I. (2003). Influence of estrogen plus progestin on breast cancer and
mammography in healthy postmenopausal women: the Women's Health Initiative
Randomized Trial. JAMA, 289(24), 3243-3253. doi:10.1001/jama.289.24.3243
51. Bassuk, S. S., & Manson, J. E. (2015). Oral contraceptives and menopausal hormone
therapy: relative and attributable risks of cardiovascular disease, cancer, and other health
outcomes. Ann Epidemiol, 25(3), 193-200. doi:10.1016/j.annepidem.2014.11.004
52. Gierisch, J. M., Coeytaux, R. R., Urrutia, R. P., Havrilesky, L. J., Moorman, P. G., Lowery,
W. J., . . . Myers, E. R. (2013). Oral contraceptive use and risk of breast, cervical,
colorectal, and endometrial cancers: a systematic review. Cancer Epidemiol Biomarkers
Prev, 22(11), 1931-1943. doi:10.1158/1055-9965.EPI-13-0298
53. White, E., Malone, K. E., Weiss, N. S., & Daling, J. R. (1994). Breast cancer among young
U.S. women in relation to oral contraceptive use. J Natl Cancer Inst, 86(7), 505-514.
doi:10.1093/jnci/86.7.505
54. Morch, L. S., Skovlund, C. W., Hannaford, P. C., Iversen, L., Fielding, S., & Lidegaard, O.
(2017). Contemporary Hormonal Contraception and the Risk of Breast Cancer. N Engl J
Med, 377(23), 2228-2239. doi:10.1056/NEJMoa1700732
19
55. Brinton, L. A., Daling, J. R., Liff, J. M., Schoenberg, J. B., Malone, K. E., Stanford, J. L., .
. . Hoover, R. N. (1995). Oral contraceptives and breast cancer risk among younger
women. J Natl Cancer Inst, 87(11), 827-835. doi:10.1093/jnci/87.11.827
56. Bjelic-Radisic, V., & Petru, E. (2010). [Hormonal contraception and breast cancer risk].
Wien Med Wochenschr, 160(19-20), 483-486. doi:10.1007/s10354-010-0807-0
57. Provenzano, E., Ulaner, G. A., & Chin, S. F. (2018). Molecular Classification of Breast
Cancer. PET Clin, 13(3), 325-338. doi:10.1016/j.cpet.2018.02.004
58. Lehmann, B. D., Jovanovic, B., Chen, X., Estrada, M. V., Johnson, K. N., Shyr, Y., . . .
Pietenpol, J. A. (2016). Refinement of Triple-Negative Breast Cancer Molecular Subtypes:
Implications for Neoadjuvant Chemotherapy Selection. PLoS One, 11(6), e0157368.
doi:10.1371/journal.pone.0157368
59. National Cancer Institute Surveillance Epidemiology and End Results Program (SEER).
(2019). Cancer Stat Facts: Female Breast Cancer Subtypes.
https://seer.cancer.gov/statfacts/html/breast-subtypes.html
60. American Cancer Society. (2019). Cancer Facts & Figures for African Americans 2019-
2021. Retrieved from Atlanta: https://www.cancer.org/content/dam/cancer-
org/research/cancer-facts-and-statistics/cancer-facts-and-figures-for-african-
americans/cancer-facts-and-figures-for-african-americans-2019-2021.pdf
61. Palmer, J. R., Boggs, D. A., Wise, L. A., Adams-Campbell, L. L., & Rosenberg, L. (2012).
Individual and neighborhood socioeconomic status in relation to breast cancer incidence in
African-American women. Am J Epidemiol, 176(12), 1141-1146. doi:10.1093/aje/kws211
62. American Cancer Society. (2019). Cancer Facts & Figures 2019. Retrieved from Atlanta:
https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-
statistics/annual-cancer-facts-and-figures/2019/cancer-facts-and-figures-2019.pdf
63. Winters, S., Martin, C., Murphy, D., & Shokar, N. K. (2017). Breast Cancer Epidemiology,
Prevention, and Screening. Prog Mol Biol Transl Sci, 151, 1-32.
doi:10.1016/bs.pmbts.2017.07.002
64. Neuhouser, M. L., Aragaki, A. K., Prentice, R. L., Manson, J. E., Chlebowski, R., Carty, C.
L., . . . Anderson, G. L. (2015). Overweight, Obesity, and Postmenopausal Invasive Breast
Cancer Risk: A Secondary Analysis of the Women's Health Initiative Randomized Clinical
Trials. JAMA Oncol, 1(5), 611-621. doi:10.1001/jamaoncol.2015.1546
65. Link, L. B., Canchola, A. J., Bernstein, L., Clarke, C. A., Stram, D. O., Ursin, G., & Horn-
Ross, P. L. (2013). Dietary patterns and breast cancer risk in the California Teachers Study
cohort. Am J Clin Nutr, 98(6), 1524-1532. doi:10.3945/ajcn.113.061184
66. Warner, E. T., Colditz, G. A., Palmer, J. R., Partridge, A. H., Rosner, B. A., & Tamimi, R.
M. (2013). Reproductive factors and risk of premenopausal breast cancer by age at
20
diagnosis: are there differences before and after age 40? Breast Cancer Res Treat, 142(1),
165-175. doi:10.1007/s10549-013-2721-9
67. Dartois, L., Fagherazzi, G., Baglietto, L., Boutron-Ruault, M. C., Delaloge, S., Mesrine, S.,
& Clavel-Chapelon, F. (2016). Proportion of premenopausal and postmenopausal breast
cancers attributable to known risk factors: Estimates from the E3N-EPIC cohort. Int J
Cancer, 138(10), 2415-2427. doi:10.1002/ijc.29987
68. Clavel-Chapelon, F., & Group, E. N. E. (2002). Differential effects of reproductive factors
on the risk of pre- and postmenopausal breast cancer. Results from a large cohort of French
women. Br J Cancer, 86(5), 723-727. doi:10.1038/sj.bjc.6600124
69. Weiderpass, E., Braaten, T., Magnusson, C., Kumle, M., Vainio, H., Lund, E., & Adami,
H. O. (2004). A prospective study of body size in different periods of life and risk of
premenopausal breast cancer. Cancer Epidemiol Biomarkers Prev, 13(7), 1121-1127.
70. Premenopausal Breast Cancer Collaborative, G., Schoemaker, M. J., Nichols, H. B.,
Wright, L. B., Brook, M. N., Jones, M. E., . . . Swerdlow, A. J. (2018). Association of
Body Mass Index and Age With Subsequent Breast Cancer Risk in Premenopausal
Women. JAMA Oncol, 4(11), e181771. doi:10.1001/jamaoncol.2018.1771
71. Colditz, G. A., Feskanich, D., Chen, W. Y., Hunter, D. J., & Willett, W. C. (2003).
Physical activity and risk of breast cancer in premenopausal women. Br J Cancer, 89(5),
847-851. doi:10.1038/sj.bjc.6601175
72. Boeke, C. E., Eliassen, A. H., Oh, H., Spiegelman, D., Willett, W. C., & Tamimi, R. M.
(2014). Adolescent physical activity in relation to breast cancer risk. Breast Cancer Res
Treat, 145(3), 715-724. doi:10.1007/s10549-014-2919-5
73. Wu, Y., Zhang, D., & Kang, S. (2013). Physical activity and risk of breast cancer: a meta-
analysis of prospective studies. Breast Cancer Res Treat, 137(3), 869-882.
doi:10.1007/s10549-012-2396-7
74. Rauscher, G. H., Campbell, R. T., Wiley, E. L., Hoskins, K., Stolley, M. R., & Warnecke,
R. B. (2016). Mediation of Racial and Ethnic Disparities in Estrogen/Progesterone
Receptor-Negative Breast Cancer by Socioeconomic Position and Reproductive Factors.
Am J Epidemiol, 183(10), 884-893. doi:10.1093/aje/kwv226
75. Bernstein, L., Allen, M., Anton-Culver, H., Deapen, D., Horn-Ross, P. L., Peel, D., . . .
Ross, R. K. (2002). High breast cancer incidence rates among California teachers: results
from the California Teachers Study (United States). Cancer Causes Control, 13(7), 625-
635.
76. Ellingjord-Dale, M., Vos, L., Tretli, S., Hofvind, S., Dos-Santos-Silva, I., & Ursin, G.
(2017). Parity, hormones and breast cancer subtypes - results from a large nested case-
control study in a national screening program. Breast Cancer Res, 19(1), 10.
doi:10.1186/s13058-016-0798-x
21
77. Ma, H., Ursin, G., Xu, X., Lee, E., Togawa, K., Duan, L., . . . Bernstein, L. (2017).
Reproductive factors and the risk of triple-negative breast cancer in white women and
African-American women: a pooled analysis. Breast Cancer Res, 19(1), 6.
doi:10.1186/s13058-016-0799-9
78. Johnson, A. M., Kirk, R., & Muzik, M. (2015). Overcoming Workplace Barriers: A Focus
Group Study Exploring African American Mothers' Needs for Workplace Breastfeeding
Support. J Hum Lact, 31(3), 425-433. doi:10.1177/0890334415573001
79. U.S. Department of Health and Human Services Health Resources and Services
Administration Maternal and Child Health Bureau. (2013). Women's Health USA 2013:
Household Composition. Retrieved from Rockville, MD:
https://mchb.hrsa.gov/whusa13/population-characteristics.html
80. Robinson, W. R., Tse, C. K., Olshan, A. F., & Troester, M. A. (2014). Body size across the
life course and risk of premenopausal and postmenopausal breast cancer in Black women,
the Carolina Breast Cancer Study, 1993-2001. Cancer Causes Control, 25(9), 1101-1117.
doi:10.1007/s10552-014-0411-5
81. van den Brandt, P. A., Ziegler, R. G., Wang, M., Hou, T., Li, R., Adami, H. O., . . . Smith-
Warner, S. A. (2021). Body size and weight change over adulthood and risk of breast
cancer by menopausal and hormone receptor status: a pooled analysis of 20 prospective
cohort studies. Eur J Epidemiol, 36(1), 37-55. doi:10.1007/s10654-020-00688-3
82. Bandera, E. V., Chandran, U., Hong, C. C., Troester, M. A., Bethea, T. N., Adams-
Campbell, L. L., . . . Rosenberg, L. (2015). Obesity, body fat distribution, and risk of breast
cancer subtypes in African American women participating in the AMBER Consortium.
Breast Cancer Res Treat, 150(3), 655-666. doi:10.1007/s10549-015-3353-z
83. Turkoz, F. P., Solak, M., Petekkaya, I., Keskin, O., Kertmen, N., Sarici, F., . . . Altundag,
K. (2013). Association between common risk factors and molecular subtypes in breast
cancer patients. Breast, 22(3), 344-350. doi:10.1016/j.breast.2012.08.005
84. Ma, H., Bernstein, L., Ross, R. K., & Ursin, G. (2006). Hormone-related risk factors for
breast cancer in women under age 50 years by estrogen and progesterone receptor status:
results from a case-control and a case-case comparison. Breast Cancer Res, 8(4), R39.
doi:10.1186/bcr1514
85. Carey, L. A., Perou, C. M., Livasy, C. A., Dressler, L. G., Cowan, D., Conway, K., . . .
Millikan, R. C. (2006). Race, breast cancer subtypes, and survival in the Carolina Breast
Cancer Study. JAMA, 295(21), 2492-2502. doi:10.1001/jama.295.21.2492
86. Tamimi, R. M., Colditz, G. A., Hazra, A., Baer, H. J., Hankinson, S. E., Rosner, B., . . .
Collins, L. C. (2012). Traditional breast cancer risk factors in relation to molecular
subtypes of breast cancer. Breast Cancer Res Treat, 131(1), 159-167. doi:10.1007/s10549-
011-1702-0
22
87. Gaudet, M. M., Gapstur, S. M., Sun, J., Diver, W. R., Hannan, L. M., & Thun, M. J.
(2013). Active smoking and breast cancer risk: original cohort data and meta-analysis. J
Natl Cancer Inst, 105(8), 515-525. doi:10.1093/jnci/djt023
88. U.S. Department of Health and Human Services. (2014). Cancer. In U.S. Department of
Health and Human Services Centers for Disease Control and Prevention National Center
for Chronic Disease Prevention and Health Promotion Office on Smoking and Health
(Ed.), The Health Consequences of Smoking-50 Years of Progress: A Report of the
Surgeon General. Atlanta, GA.
89. Xue, F., Willett, W. C., Rosner, B. A., Hankinson, S. E., & Michels, K. B. (2011).
Cigarette smoking and the incidence of breast cancer. Arch Intern Med, 171(2), 125-133.
doi:10.1001/archinternmed.2010.503
90. Bao, Y., Bertoia, M. L., Lenart, E. B., Stampfer, M. J., Willett, W. C., Speizer, F. E., &
Chavarro, J. E. (2016). Origin, Methods, and Evolution of the Three Nurses' Health
Studies. Am J Public Health, 106(9), 1573-1581. doi:10.2105/AJPH.2016.303338
91. Calle, E. E., Rodriguez, C., Jacobs, E. J., Almon, M. L., Chao, A., McCullough, M. L., . . .
Thun, M. J. (2002). The American Cancer Society Cancer Prevention Study II Nutrition
Cohort: rationale, study design, and baseline characteristics. Cancer, 94(2), 500-511.
doi:10.1002/cncr.10197
92. Rosenberg, L., Boggs, D. A., Bethea, T. N., Wise, L. A., Adams-Campbell, L. L., &
Palmer, J. R. (2013). A prospective study of smoking and breast cancer risk among
African-American women. Cancer Causes Control, 24(12), 2207-2215.
doi:10.1007/s10552-013-0298-6
93. White, A. J., D'Aloisio, A. A., Nichols, H. B., DeRoo, L. A., & Sandler, D. P. (2017).
Breast cancer and exposure to tobacco smoke during potential windows of susceptibility.
Cancer Causes Control, 28(7), 667-675. doi:10.1007/s10552-017-0903-1
94. Reynolds, P., Goldberg, D., Hurley, S., Nelson, D. O., Largent, J., Henderson, K. D., &
Bernstein, L. (2009). Passive smoking and risk of breast cancer in the California teachers
study. Cancer Epidemiol Biomarkers Prev, 18(12), 3389-3398. doi:10.1158/1055-
9965.EPI-09-0936
95. Marcus, P. M., Newman, B., Millikan, R. C., Moorman, P. G., Baird, D. D., & Qaqish, B.
(2000). The associations of adolescent cigarette smoking, alcoholic beverage consumption,
environmental tobacco smoke, and ionizing radiation with subsequent breast cancer risk
(United States). Cancer Causes Control, 11(3), 271-278. doi:10.1023/a:1008911902994
96. Chuang, S. C., Gallo, V., Michaud, D., Overvad, K., Tjonneland, A., Clavel-Chapelon, F., .
. . Vineis, P. (2011). Exposure to environmental tobacco smoke in childhood and incidence
of cancer in adulthood in never smokers in the European Prospective Investigation into
Cancer and Nutrition. Cancer Causes Control, 22(3), 487-494. doi:10.1007/s10552-010-
9723-2
23
97. Bonner, M. R., Nie, J., Han, D., Vena, J. E., Rogerson, P., Muti, P., . . . Freudenheim, J. L.
(2005). Secondhand smoke exposure in early life and the risk of breast cancer among never
smokers (United States). Cancer Causes Control, 16(6), 683-689. doi:10.1007/s10552-
005-1906-x
98. Weiss, H. A., Potischman, N. A., Brinton, L. A., Brogan, D., Coates, R. J., Gammon, M.
D., . . . Schoenberg, J. B. (1997). Prenatal and perinatal risk factors for breast cancer in
young women. Epidemiology, 8(2), 181-187. doi:10.1097/00001648-199703000-00010
99. Sanderson, M., Williams, M. A., Malone, K. E., Stanford, J. L., Emanuel, I., White, E., &
Daling, J. R. (1996). Perinatal factors and risk of breast cancer. Epidemiology, 7(1), 34-37.
doi:10.1097/00001648-199601000-00007
100. Sanderson, M., Williams, M. A., Daling, J. R., Holt, V. L., Malone, K. E., Self, S. G., &
Moore, D. E. (1998). Maternal factors and breast cancer risk among young women.
Paediatr Perinat Epidemiol, 12(4), 397-407. doi:10.1046/j.1365-3016.1998.00133.x
101. Potischman, N., & Troisi, R. (1999). In-utero and early life exposures in relation to risk of
breast cancer. Cancer Causes Control, 10(6), 561-573. doi:10.1023/a:1008955110868
102. Manjer, J., Malina, J., Berglund, G., Bondeson, L., Garne, J. P., & Janzon, L. (2001).
Smoking associated with hormone receptor negative breast cancer. Int J Cancer, 91(4),
580-584.
103. Kawai, M., Malone, K. E., Tang, M. T., & Li, C. I. (2014). Active smoking and the risk of
estrogen receptor-positive and triple-negative breast cancer among women ages 20 to 44
years. Cancer, 120(7), 1026-1034. doi:10.1002/cncr.28402
104. Centers for Disease Control and Prevention Office on Smoking and Health. (2019).
Smoking & Tobacco Use: African Americans and Tobacco Use.
https://www.cdc.gov/tobacco/disparities/african-americans/index.htm
105. Kypriotakis, G., Robinson, J. D., Green, C. E., & Cinciripini, P. M. (2018). Patterns of
Tobacco Product Use and Correlates Among Adults in the Population Assessment of
Tobacco and Health (PATH) Study: A Latent Class Analysis. Nicotine Tob Res,
20(suppl_1), S81-S87. doi:10.1093/ntr/nty025
106. Perez-Stable, E. J., Herrera, B., Jacob, P., 3rd, & Benowitz, N. L. (1998). Nicotine
metabolism and intake in black and white smokers. JAMA, 280(2), 152-156.
doi:10.1001/jama.280.2.152
107. Ross, K. C., Gubner, N. R., Tyndale, R. F., Hawk, L. W., Jr., Lerman, C., George, T. P., . .
. Benowitz, N. L. (2016). Racial differences in the relationship between rate of nicotine
metabolism and nicotine intake from cigarette smoking. Pharmacol Biochem Behav, 148,
1-7. doi:10.1016/j.pbb.2016.05.002
108. Hung, M. C., Ekwueme, D. U., Rim, S. H., & White, A. (2016). Racial/ethnicity disparities
in invasive breast cancer among younger and older women: An analysis using multiple
24
measures of population health. Cancer Epidemiol, 45, 112-118.
doi:10.1016/j.canep.2016.10.013

25
Table 1.1 Summary of breast cancer susceptibility genes of high and moderate penetrance
Gene Penetrance Function
Related
Syndromes /
Conditions
Estimated
Contribution
to Breast
Cancer (%)
BRCA1 High
Tumor suppressor; involved in genomic
stability, DNA repair, cell cycle
regulation, chromatin remodeling,
transcriptional regulation, and protein
activity regulation by ubiquitylation
Hereditary
breast and
ovarian
cancer
1-4
BRCA2 High
Homologous recombination in meiosis and
in repair of double-strand breaks

1-2
CDH1 High
Encodes for E-cadherin, which is involved
in cell-to-cell adhesion
Familial diffuse
gastric cancer
<1
PTEN High
Tumor suppressor and cell growth
regulator; down-regulates the
phosphatidylinositol-3-kinase (PI3K)
signal transduction cascade
Cowden
syndrome
<1
STK11 High
Tumor suppressor; Encodes a
serine/threonine kinase
Peutz–Jeghers
syndrome
<1
TP53 High
Tumor suppressor; codes for p53 protein
which can induce cell cycle arrest and
apoptosis
Li–Fraumeni
syndrome <1
ATM Moderate
Encodes a checkpoint kinase which
phosphorylates the p53 and BRCA1
proteins
Ataxia-
telangiectasia <1-2
BRIP1 Moderate
Encodes for C-terminal helicase 1 which
interacts with BRCA1
Fanconi anemia
<1
CHEK2 Moderate
Encodes a serine/threonine checkpoint
kinase for cell-cycle control and DNA
repair

1-2
NBN Moderate
Part of the MRN complex involved in the
recognition and repair of DNA double-
strand breaks
Nijmegen
breakage
syndrome
<1
PALB2 Moderate
Homologous recombination in repair of
double-strand breaks

<1
References: 19; 40; 41; 42

26
Table 1.2 Summary of risk factors by breast cancer subtype
Breast Cancer
Subtype
Risk factor
Direction of the
association
Luminal A Age at menarche
Parity
Age at first birth
Age of ≥40 years
Alcohol consumption
Estrogen plus progestin hormone use
Having two first-degree relatives with breast cancer
Postmenopausal hormone use
-
-
+
+
+
+
+
+
Luminal B Parity
Weight gain since age 18
+
+
HER2-enriched Age at menopause +
TNBC Breastfeeding
BRCA1 mutations (associated with basal-like TNBC)
Current estrogen plus progestin hormone use
Obesity
Parity
-
+
+
+
+
References: 57; 76; 83; 85; 86

27
Figure 1.1 Summary of the pathways of estrogen-related carcinogenesis

Estrogen receptor activation
1. Activates genomic, non-
genomic and/or mitochondrial
processes
2. Altered gene expression
3. Increased cell proliferation;
decreased apoptosis
Epigenetic effects of estrogen
1. Activates genomic processes
2. Altered gene expression
3. Increased cell proliferation;
decreased apoptosis
Estrogen metabolism

Pathway 1
1. Oxidized to 16a-
hydroxyestrone
2. Forms covalent bonds to
proteins and DNA

Pathway 2
1. Oxidized to 2- and 4-
hydroxyestrone or
hydroxyestradiol
2. Metabolized to quinones
3. Adducts form; oxidative
DNA damage
Adapted from Yager. Steroids. 2015.
28
Figure 1.2 Breast cancer incidence by subtype among non-Hispanic (NH) White and NH Black
women under 50 years of age, SEER 21 2010-2016

0
100
200
300
400
2010 2011 2012 2013 2014 2015 2016
Incidence (per 1,000,000)
Year of Diagnosis
SEER 21 incidence by BC subtype
among NH Black and NH White
women age <50 years, 2010-2016
HR+/HER2- (Luminal A)
HR+/HER2+ (Luminal B)
HR-/HER2+ (HER2-enriched)
HR-/HER2- (TNBC)
Unknown
0
50
100
150
200
250
300
350
2010 2011 2012 2013 2014 2015 2016
Incidence (per 1,000,000)
Year of Diagnosis
NH Black
0
50
100
150
200
250
300
350
2010 2011 2012 2013 2014 2015 2016
Incidence (per 1,000,000)
Year of Diagnosis
NH White
29
CHAPTER TWO: A Systematic Review and Meta-Analysis of Smoking and Circulating
Sex-Hormone Levels Among Premenopausal Women

Authors: Ugonna Ihenacho, MPH
1
, Intira Sriprasert, MD,

PhD
1
, Wendy J. Mack, PhD
1
, Ann S.
Hamilton, PhD
1
, Jennifer B. Unger, PhD
1
, Michael F. Press, MD, PHD
2
, and Anna H. Wu, PhD
1*

Institutional Affiliations:
1
Department of Preventive Medicine, Keck School of Medicine of
USC;
2
Department of Pathology, Keck School of Medicine of USC

2.1 Abstract
Background: It is well established that pre-diagnostic circulating androgen and estrogen levels
are associated with risk of breast cancer development in premenopausal and postmenopausal
women. Pooled analyses in postmenopausal women found higher androgen and estrogen levels
in current heavy cigarette smokers compared to nonsmokers. However, evidence among
premenopausal women has been inconsistent.
Methods: We conducted a systematic PubMed review and collated all published articles on
urinary or blood sex hormone levels by smoking status among healthy, premenopausal women
who were nonusers of exogenous hormones. A random effects meta-analysis was conducted to
combine the standardized mean differences (SMD) and 95% confidence intervals (CIs) for
estradiol, progesterone, testosterone, dehydroepiandrosterone (DHEA), DHEA-sulfate
(DHEAS), and sex hormone-binding globulin (SHBG) by smoking status. Findings were
summarized by menstrual cycle phase (follicular, luteal, or varied) and overall.
30
Results: Mean sex hormone levels from 19 published peer-reviewed articles were included.
Compared to nonsmokers, current smokers showed elevated levels of follicular phase estradiol
and progesterone and lower levels of luteal phase estradiol. Significantly increased testosterone
levels among smokers compared to nonsmokers were identified from studies with varied
menstrual phase timing (SMD 0.14; 95% CI 0.0005, 0.29) and for DHEAS levels over all phases
(SMD 0.12; 95% CI 0.01, 0.22). Substantial heterogeneity existed in these studies.
Conclusions: This meta-analysis of cross-sectional data suggests that smoking may adversely
alter blood hormone levels in healthy premenopausal women; however, the differences were
small. Larger and covariate-adjusted studies are needed to better understand this relationship and
to reduce heterogeneity.

2.2 Introduction
Current evidence suggests that high levels of pre-diagnostic circulating androgen and
estrogen levels are associated with an increased risk of breast cancer among premenopausal and
postmenopausal women (1-3). High endogenous estradiol (E2) concentrations may contribute to
breast carcinogenesis through various pathways. In one pathway, E2 is metabolized to catechol
estrogens and other hydroxylated metabolites that form deoxyribonucleic acid (DNA) adducts
and cause oxidative DNA damage (1). Alternatively, estrogen receptor activation is implicated in
triggering signaling pathways that lead to altered gene expression, impaired regulation of breast
cell proliferation, and dysfunctional apoptosis (1). These carcinogenic pathways can also be
activated by testosterone (T), androstenedione and other androgens that can be converted to
estrone (E1) and E2. Specifically, dehydroepiandrosterone (DHEA) and DHEA-sulfate (DHEAS)
can be converted to T or E2 through the activation of the hydroxysteroid dehydrogenase and
31
aromatase enzymes. Also, sex hormone-binding globulin (SHBG) regulates levels of free E2 and
T (1; 4; 5).
Although demographic and lifestyle factors, including alcohol consumption and physical
activity, have been associated with endogenous sex hormone levels in both premenopausal and
postmenopausal populations, results on smoking and sex hormonal levels have been less
consistent (2; 6-8). In a pooled analysis of 6291 healthy postmenopausal women, compared to
never smoking, current heavy smoking of 15 or more cigarettes per day was significantly
associated with an approximate 13% increase in mean E2 and free E2 levels, a 15% increase in
mean DHEAS levels, and a 20% and 19% increase in mean levels of T and free T, respectively
(6). However, among current smokers of less than 15 cigarettes per day, mean E2 and free E2
levels were 12% and 11% lower than nonsmokers (6). A few studies have described differences
in hormone levels by smoking status among premenopausal women, but the effect sizes were
often small or lacked statistical significance (2; 9; 10).
We hypothesized that smoking is associated with mean E2, progesterone, androgen (T,
DHEA, DHEAS) and SHBG levels in premenopausal women. We conducted a systematic
review and meta-analysis of publications to estimate the difference in standardized mean
hormone levels among current premenopausal smokers compared to nonsmokers.

2.3 Methods
2.3.1 Data source and search terms
Articles were collected from PubMed using a list of search terms developed by two
reviewers (UI, IS). Observational studies reporting hormone levels by smoking status through
December 2019 were reviewed. Medical Search Heading (MeSH) terms for smoking terms
32
(“tobacco smoking" or "cigarette smoking” or “smoking") and research population parameters
(“Female” not "Mice") were used (Supplementary Table 2.1). Searches were conducted using
these terms and sex hormones of interest ("Gonadal Steroid Hormones" "Estrogens"
"Progesterone" "Androgens" "Testosterone" "Dehydroepiandrosterone"
"Dehydroepiandrosterone Sulfate" "Sex Hormone-Binding Globulin"). Results were compiled in
the Covidence Systematic Review Software (Covidence) for screening; duplicate articles were
automatically removed by the software (Veritas Health Innovation, Melbourne, Australia.
www.covidence.org). The title and abstracts of the compiled articles were evaluated to identify
articles for exclusion from full text screening. The remaining articles were read to identify
publications meeting criteria for data extraction. We reviewed the references cited in the selected
articles to identify additional studies. The reviewers identified articles for inclusion by consensus
or by referring to a third reviewer (WM). A review protocol and registration were not needed for
this study since it was an evaluation of published data and did not require Institutional Review
Board approval.

2.3.2 Study selection
We included observational studies that measured blood or urinary hormone levels by
personal smoking status in populations of healthy, premenopausal, adult women who were
nonusers of exogenous hormones, including oral contraceptives (OC). To maximize use of
available data, we did not exclude studies based on numbers of participants or date of
publication. We excluded studies of perimenopausal women, pregnant women, women with
diagnosed hormonal-related conditions (e.g., polycystic ovary syndrome), or studies that did not
exclude hormone users or stratify on exogenous hormone use. Articles published in languages
33
other than English, unpublished reports, abstract-only studies, letters, and review articles were
excluded.

2.3.4 Data extraction
Information extracted from the selected full-text articles included the publication details
(authors, title, year of publication and journal), study design (cross-sectional, cohort or case-
control), characteristics of the study population (inclusion and exclusion criteria), region of
origin (Europe, North America, International), sample type (blood or urine), menstrual phase of
the sample collection (follicular, luteal, varied), smoking status (nonsmoker, never smoker, ever
smoker, current smoker), number of participants by smoking status, the mean, geometric mean or
median hormone level and corresponding measures of variability (standard deviation (SD),
standard error (SE), 95% confidence interval (CI), interquartile range (IQR), and range) or P-
value, and the covariates included in adjusted analysis. Two attempts to contact the authors of a
study were made if measures of variability, a P-value for differences or any other summary data
were not available in the article.

2.3.5 Data management and analysis
The meta-analysis was conducted with data on an arithmetic scale. Studies reporting
geometric mean estimates were transformed to the arithmetic scale using procedures described
by Higgins et.al. (11). An SD was estimated from a mean and P-value using methods outlined in
the Cochrane Handbook for Systematic Reviews of Interventions (12). We estimated means and
SDs from medians and IQRs using established formulas (Supplementary Table 2.2) (13).
Smoking status was dichotomized as “smoker” for current smokers and “nonsmoker” for never
34
and nonsmoker, as specified in the publication. Former smokers were omitted from the analysis
whenever provided as a separate smoking status category. We calculated the standardized mean
differences (SMD) and corresponding 95% CI and used these to compute a summary effect
estimate using a random-effects meta-analysis. Forest plots were generated to depict the
contribution of each study and resulting summary estimate. We used the I
2
statistic to evaluate
the heterogeneity of smoking effect estimates in the sample. Interpretations of heterogeneity used
published guidelines (12): minimal heterogeneity as an I
2
of 0-29%; moderate heterogeneity as
an I
2
of 30-59%; and I
2
of 60-100% for substantial heterogeneity. Funnel plots were generated
for the visual assessment of publication bias; Begg’s and Egger’s tests were evaluated for a
quantitative assessment of publication bias. To examine heterogeneity in smoking effects among
studies, we conducted a residual maximum likelihood meta-regression to estimate the impact of
the year of publication, use of adjusted (versus crude) effect estimates, and geographic study
region (Europe vs. North America) on the pooled sample estimates overall and by menstrual
phase. We also conducted several sensitivity analyses. We conducted a leave-one-out evaluation
where analysis was conducted with each study removed separately. We conducted subgroup
analyses, limiting to: (i) studies that reported covariate-adjusted smoking effect estimates and,
(ii) results based on blood samples only. Each sensitivity analysis was conducted overall and by
menstrual phase.
We used the observed effect sizes from the E2 analysis to develop curves to estimate the
power detected for various sample sizes using the SMD and number of studies (k) observed in
the meta-analysis. The curves were developed for the analysis of E2, overall and by menstrual
phase, since this was the analyte with the largest sample size.
35
Reporting guidelines outlined by the Meta-Analyses of Observational Studies in
Epidemiology (MOOSE) and Preferred Reporting Items for Systematic Reviews and Meta-
Analyses (PRISMA) were followed (14; 15). All statistical analyses were conducted in Stata
14.2 (StataCorp, College Station, TX).

2.4 Results
2.4.1 Literature review and study characteristics
Figure 2.1 shows the details of the PubMed article selection process which yielded 1,325
studies. After removing duplicates, 864 articles remained for title and abstract review, with 81
selected for full-text review. We excluded 64 of the 81 studies for various reasons: a study
sample comprised of women who were perimenopausal, pregnant, had a diagnosed hormone-
related condition, or other population-level characteristics that did not meet inclusion criteria (n
= 19); no report of hormone level by smoking status (n = 16); women using OCs or other
exogenous hormones or use not specified (n = 11); an intervention or review article (n = 6); the
article was not in English (n = 2); and the same study population was included in another article
that was selected for inclusion (n = 5). Information on the sample size, effect estimate, or other
summary data were not available in five studies; at least two attempts were made to contact the
corresponding author before they were excluded. We identified two additional articles from a
review of bibliographic references. In total, 19 articles published over 20 years were included
from the systematic review. E2 levels by smoking status were provided in 16 articles;
progesterone levels in nine; T in nine; DHEA in two; DHEAS in three; and SHBG in nine (2; 8-
10; 16-30).
36
The characteristics of the 19 articles are presented by hormone in Table 2.1. In total, the
mean sex hormone levels of 4,531 smokers and 12,938 nonsmokers were abstracted. The
hormone measures were estimated largely from blood samples (91.1%) that were collected in the
follicular menstrual phase (50.0%). Most of the studies were cross-sectional (78.9%) by design;
four were cross-sectional analyses of hormone levels and smoking status from cohort studies.

2.4.2 Association between smoking and hormone levels
Results for smoking and hormone levels by menstrual phase and overall are summarized
in Table 2.2 and the resulting forest and funnel plots are presented in Figures 2.2 and 2.3. E2
levels did not differ significantly by smoking status overall (SMD -0.03; 95% CI -0.18, 0.13)
(Table 2.2, Figure 2.2a). For estimates measured in the follicular phase and at varied timepoints
in the menstrual phase, E2 levels were higher among smokers than nonsmokers (follicular SMD
0.10; 95% CI -0.17, 0.38; varied cycle SMD 0.07; 95% CI -0.18, 0.03). Conversely, E2 levels in
the luteal phase were lower among smokers (SMD -0.17; 95% CI -0.53, 0.19). The estimates did
not differ significantly by menstrual phase. The funnel plot (Figure 2.3a) shows that the smaller
studies contributed to a positive effect and a single study in the luteal phase contributed to a
positive bias.
Progesterone levels were non-significantly higher in smokers compared to nonsmokers
overall, in the follicular phase, and in the luteal phase (overall SMD 0.12; 95% CI -0.12, 0.36;
follicular SMD 0.27; 95% CI -0.17, 0.72; luteal SMD 0.02; 95% CI -0.28, 0.32) (Table 2.2,
Figure 2.2b). One study in the luteal phase contributed to a negative bias (Figure 2.3b).
Level of T was non-significantly increased among smokers compared to nonsmokers
overall, for measures at varied timepoints and among a single study in the luteal phase (overall
37
SMD 0.14; 95% CI -0.03, 0.30) (Table 2.2, Figure 2.2c). A significant difference in levels was
observed among the measures with varied sample collection (SMD 0.14; 95% CI 0.0005, 0.29)
but there was some evidence of publication bias by Egger’s test in the subsample (P = 0.02, data
not shown).
Based on three studies, a significantly higher mean DHEAS level was observed among
smokers (overall SMD 0.12; 95% CI 0.01, 0.22) (Table 2.2, Figure 2.2e). Smokers had a higher
mean DHEA level compared to nonsmokers in the follicular, luteal, and overall phases (overall
SMD 0.76; 95% CI -0.45, 1.97). Lastly, SHBG levels among smokers were elevated in the
follicular phase (SMD 0.15; -0.35, 0.65) (Figure 2.2f) but lower in the luteal phase (SMD -0.19; -
0.91, 0.53) thus contributing to a nearly negligible SMD overall (SMD -0.01; 95% CI -0.15,
0.13).

2.4.3 Assessment of heterogeneity and sensitivity analyses
Heterogeneity between studies varied for each of the hormones. We observed substantial
heterogeneity (I
2
range: 60%-100%)

in the smoking-related effect estimates on E2, progesterone
and DHEA measures assessed overall and in each menstrual phase (Table 2.2). Substantial
heterogeneity was observed for T measures in the follicular phase (I
2
=61.1%) while
heterogeneity was modest for T measures with varied collection timing (31.4%) and overall
(48.9%). Minimal heterogeneity was observed in the estimates for DHEAS in the follicular phase
and overall (two studies and three studies included, respectively; I
2
= 0% for both phases).
Substantial heterogeneity was observed in the follicular and luteal phases for the SHBG
measures, but moderate heterogeneity was observed for SHBG measures with varied collection
timing and overall.
38
In the meta-regression, year of publication and reporting of covariate-adjusted hormone
measures did not significantly contribute to heterogeneity, overall or by menstrual phase (data
not shown). A few changes in results emerged in the leave-one-out sensitivity analysis.
Removing the Daniel study (20) resulted in a significant difference in T levels by smoking status
(overall SMD 0.17; 95% CI 0.01, 0.33). After removing the Soldin study (9), a significant
difference in DHEA levels was observed (overall SMD 1.31; 95% CI 0.27, 2.35). Omitting the
Endogenous Hormones Breast Cancer Collaborative Group (EHBCCG) study (2), measures with
varied timing resulted in a significant difference in T (varied SMD 0.26; 95% CI 0.01, 0.50) and
SHBG levels by smoking status (varied SMD -0.19; 95% CI -0.36, -0.02).
In the sensitivity analysis limited to studies with covariate-adjusted estimates (48% of the
measures available; Supplementary Table 2.3), smoking was associated with a 0.17 SD (95% CI
-0.75, 0.40) mean decrease in E2 levels in the follicular phase (Supplementary Figure 2.1a). A
significant increase in progesterone levels (SMD 1.31; 95% CI 0.34, 2.29) in the follicular phase
was based on the findings from only one study (Supplementary Figure 2.1b). We also found
higher DHEA levels among smokers compared to nonsmokers, overall (Supplementary Figure
2.1d; SMD 1.31; 95% CI 0.27, 2.35).
The sensitivity analysis limited to studies with blood samples was conducted for the
model estimating E2 levels because it was the only analyte with more than one study with urine
hormone measures (other hormone summaries were based on blood samples only). In these
assessments, five measures from urine samples were removed: two from the follicular phase and
three from the luteal phase. The SMD in E2 level in the follicular phase approached the null in
the sensitivity analysis (SMD 0.04; 95% CI -0.26, 0.34, data not shown) compared to the SMD
of 0.10 SDs that was observed in the full analysis. In the luteal phase, the estimated SMD in E2
39
level was higher in the sensitivity analysis compared to the full analysis (SMD 0.29; 95% CI -
1.54, 2.12). These findings in the sensitivity analysis remained nonsignificant.

2.4.4 Power analysis for estradiol measures
We estimated the power for the hormone with the largest sample size, E2, using the
observed effect sizes from the meta-analysis (Figure 2.4). If the SMDs estimated in this analysis
reflected the true effect of smoking on hormone levels, the observed SMD of -0.03 SDs for
overall E2 levels was underpowered (power=38.6%) to detect a significant difference by smoking
status in a sample of 19 studies with substantial heterogeneity. The sample of 10 studies with
substantial heterogeneity that contributed E2 estimates in the follicular phase was underpowered
(67.5%) to detect a significant SMD of 0.10. Although only 6 studies contributed to the analysis
in the luteal phase, the sample was sufficiently powered to detect differences at an effect size of -
0.17 SDs (power=92.0%). The three studies with E2 assessments with varied sample timing had
minimal heterogeneity but were underpowered to detect a significant SMD of -0.07
(power=52.2%).

2.5 Discussion
2.5.1 Summary of findings
In this systematic review and meta-analysis, we selected 19 articles published since 1988
to evaluate the association between smoking and endogenous levels of E2, progesterone, T,
DHEA, DHEAS and SHBG. Levels of four of the six hormones (progesterone, T, DHEA,
DHEAS) were higher in current smokers than nonsmokers (SMDs ranged from 0.12 to 0.76);
only DHEAS, based on three studies, was statistically significant (overall SMD 0.12; 95% 0.01,
40
0.22). Smokers and nonsmokers showed little difference in E2 (-0.03) and SHBG levels (-0.01)
but there were suggestive differences by menstrual phase, showing higher levels among smokers
in the follicular phase but lower levels among smokers in the luteal phase. The summary SMDs
in the covariate-adjusted sensitivity analysis suggested larger effects of smoking than in the full
meta-analysis; however, the positive association in the follicular phase for progesterone (SMD
1.31, 95% CI 0.34, 2.29) and negative association for DHEA in the luteal phase (SMD 1.90, 95%
CI 0.76, 3.04) were each based on only one study.

2.5.2 Improving power for estradiol measures
Overall, this analysis was underpowered to detect a smoking-related difference in mean
E2 levels in the presence of substantial heterogeneity (power=38.6%). However, with moderate
heterogeneity and a larger population (i.e., ~1.5 times the observed sample), power would
improve to ~80% to detect a significant difference, if one exists. However, among follicular E2
samples, sufficient power would have been achieved with the current sample size if it had been
conducted among studies with moderate heterogeneity instead of with substantial heterogeneity.
The assessments with varied sample timing had minimal heterogeneity, but a sample twice that
of the observed analysis would have been required to detect differences at an effect size of -0.07
SDs with sufficient power. The nearly null effect of smoking on overall E2 reflected the likely
chance finding of opposing effects of smoking by menstrual phase since we are not aware of a
biological reason. With larger samples for assessments by menstrual phase, it may have been
possible to develop sufficiently powered studies to identify significant differences in hormone
levels if they truly exist.

41
2.5.3 Known breast cancer risk factors and hormone levels in premenopausal women
A pooled analysis of seven prospective studies was conducted by the EHBCCG to
evaluate the association between sex hormone levels and breast cancer risk in premenopausal
populations (2). Previous studies have found that increasing alcohol intake was associated with
increased T levels in premenopausal and postmenopausal populations (2; 6; 31). The positive
association between alcohol consumption and breast cancer risk in premenopausal and
postmenopausal women is also well-established (32-35). While alcohol consumption and
cigarette smoking are highly correlated, alcohol use was only adjusted for in one study in our
meta-analysis (29; 34).
High physical activity is a protective factor for breast cancer and was associated with
lower free luteal E2 and free T levels in the Nurses’ Health Study cohort (36). In a recent UK
Biobank study, higher levels of self-reported and accelerometer-measured physical activity were
associated with lower free E2, total T, and free T levels (37). However, only one study on
smoking and hormone levels included in our meta-analysis adjusted for physical activity (29).
Smoking may be associated with small changes in hormone levels, conferring significant
biological impact but these relatively small effects are difficult to detect in the absence of large
study populations or a wide distribution of smoking exposure including heavy levels of exposure.
A lack of covariate-adjusted measures of association and the imprecise timing of sample
collections additionally contribute to residual confounding and errors in hormonal measures
among premenopausal populations.

2.5.4 Strengths and limitations
42
This study used published literature to conduct this analysis, but our approach was
comprehensive, considering carefully menstrual phase (follicular/luteal) while still allowing for
some variation in sample collection timing (varied), and included several sensitivity analyses,
adding substantially to previous studies.
There was substantial heterogeneity in many of the analyses. Neither study region nor use
of covariate-adjusted estimates explained the heterogeneity. Heterogeneity may be due to other
unmeasured variations that affected the study quality; for example, menstrual cycle day and
participant age were accounted for only in several studies. Although alcohol intake, physical
activity and BMI are associated with hormone levels, and are independently associated with
smoking (38-41), body size measures were controlled for in only 5 of the 19 studies (2; 20; 23;
29; 30) while alcohol use and physical activity were only considered in one study (29). Potential
residual confounding bias in the summary estimates could not be ruled out. Our sensitivity
analysis demonstrated that using covariate-adjusted estimates may have allowed us to estimate
larger effect sizes. Treatment of former smokers was not uniform in previous studies. Former
smokers that had not smoked for a specified period were considered as non-smokers in some
studies (9; 18; 20; 22; 27; 30) or other studies did not clearly describe whether former smokers
were excluded (16; 17; 21; 24; 25). Hence, differential misclassification of the tobacco smoke
exposure cannot be ruled out, although hormone levels are suggested to return to near never-
smoker levels 1-2 years after smoking cessation in postmenopausal women (42). With additional
well-designed studies with covariate-adjusted hormone levels and the exclusion of former
smokers, it may be possible to identify additional differences in hormone levels by smoking
status.

43
2.5.6 Future directions
We identified elevated levels of follicular E2 and progesterone, and higher luteal T levels
among premenopausal smokers compared to nonsmokers. These observations correspond with
differences in hormone levels that have been associated with increased breast cancer risk in prior
studies. However, moderate and substantial heterogeneity was observed for many of the
summary estimates. Adequately powered studies allowing for adjustment of relevant variables
such as BMI, physical activity, and alcohol use, are encouraged as this will lead to improved
assessment of the relationship between smoking and endogenous hormone levels (E2,
progesterone, T, DHEA, DHEAS, SHBG and their metabolites), overall and by menstrual phase,
among healthy premenopausal women.

44
2.6 References
1. Yager, J. D., & Davidson, N. E. (2006). Estrogen carcinogenesis in breast cancer. N Engl J
Med, 354(3), 270-282. doi:10.1056/NEJMra050776
2. Endogenous Hormones Breast Cancer Collaborative Group, Key, T. J., Appleby, P. N.,
Reeves, G. K., Travis, R. C., Alberg, A. J., . . . Vineis, P. (2013). Sex hormones and risk of
breast cancer in premenopausal women: a collaborative reanalysis of individual participant
data from seven prospective studies. Lancet Oncol, 14(10), 1009-1019. doi:10.1016/S1470-
2045(13)70301-2
3. Key, T., Appleby, P., Barnes, I., Reeves, G., Endogenous, H., & Breast Cancer
Collaborative, G. (2002). Endogenous sex hormones and breast cancer in postmenopausal
women: reanalysis of nine prospective studies. J Natl Cancer Inst, 94(8), 606-616.
doi:10.1093/jnci/94.8.606
4. Prough, R. A., Clark, B. J., & Klinge, C. M. (2016). Novel mechanisms for DHEA action.
J Mol Endocrinol, 56(3), R139-155. doi:10.1530/JME-16-0013
5. Fortunati, N., Catalano, M. G., Boccuzzi, G., & Frairia, R. (2010). Sex Hormone-Binding
Globulin (SHBG), estradiol and breast cancer. Mol Cell Endocrinol, 316(1), 86-92.
doi:10.1016/j.mce.2009.09.012
6. Endogenous Hormones Breast Cancer Collaborative Group, Key, T. J., Appleby, P. N.,
Reeves, G. K., Roddam, A. W., Helzlsouer, K. J., . . . Strickler, H. D. (2011). Circulating
sex hormones and breast cancer risk factors in postmenopausal women: reanalysis of 13
studies. Br J Cancer, 105(5), 709-722. doi:10.1038/bjc.2011.254
7. Sowers, M. F., Beebe, J. L., McConnell, D., Randolph, J., & Jannausch, M. (2001).
Testosterone concentrations in women aged 25-50 years: associations with lifestyle, body
composition, and ovarian status. Am J Epidemiol, 153(3), 256-264.
8. Barbieri, R. L., Sluss, P. M., Powers, R. D., McShane, P. M., Vitonis, A., Ginsburg, E., &
Cramer, D. C. (2005). Association of body mass index, age, and cigarette smoking with
serum testosterone levels in cycling women undergoing in vitro fertilization. Fertility and
sterility, 83(2), 302-308. doi:10.1016/j.fertnstert.2004.07.956
9. Soldin, O. P., Makambi, K. H., Soldin, S. J., & O'Mara, D. M. (2011). Steroid hormone
levels associated with passive and active smoking. Steroids, 76(7), 653-659.
doi:10.1016/j.steroids.2011.02.042
10. Dafopoulos, A., Dafopoulos, K., Georgoulias, P., Galazios, G., Limberis, V., Tsikouras, P.,
. . . Maroulis, G. (2010). Smoking and AMH levels in women with normal reproductive
history. Arch Gynecol Obstet, 282(2), 215-219. doi:10.1007/s00404-010-1425-1
11. Higgins, J. P., White, I. R., & Anzures-Cabrera, J. (2008). Meta-analysis of skewed data:
combining results reported on log-transformed or raw scales. Stat Med, 27(29), 6072-6092.
doi:10.1002/sim.3427
45
12. Higgins, J. P. T., Thomas, J., Chandler, J., Cumpston, M., Li, T., Page, M. J., & Welch, V.
A. e. (2019). Cochrane Handbook for Systematic Reviews of Interventions. from Cochrane
www.training.cochrane.org/handbook
13. Wan, X., Wang, W., Liu, J., & Tong, T. (2014). Estimating the sample mean and standard
deviation from the sample size, median, range and/or interquartile range. BMC Med Res
Methodol, 14, 135. doi:10.1186/1471-2288-14-135
14. Stroup, D. F., Berlin, J. A., Morton, S. C., Olkin, I., Williamson, G. D., Rennie, D., . . .
Thacker, S. B. (2000). Meta-analysis of observational studies in epidemiology: a proposal
for reporting. Meta-analysis Of Observational Studies in Epidemiology (MOOSE) group.
Jama, 283(15), 2008-2012. doi:10.1001/jama.283.15.2008
15. Moher, D., Liberati, A., Tetzlaff, J., Altman, D. G., & Group, P. (2009). Preferred
reporting items for systematic reviews and meta-analyses: the PRISMA statement. PLoS
Med, 6(7), e1000097-e1000097. doi:10.1371/journal.pmed.1000097
16. Longcope, C., & Johnston, C. C. (1988). Androgen and estrogen dynamics in pre- and
postmenopausal women: a comparison between smokers and nonsmokers. The Journal of
clinical endocrinology and metabolism, 67(2), 379-383. doi:10.1210/jcem-67-2-379
17. Michnovicz, J. J., Naganuma, H., Hershcopf, R. J., Bradlow, H. L., & Fishman, J. (1988).
Increased urinary catechol estrogen excretion in female smokers. Steroids, 52(1-2), 69-83.
doi:10.1016/0039-128x(88)90218-8
18. Zumoff, B., Miller, L., Levit, C. D., Miller, E. H., Heinz, U., Kalin, M., . . . Rosenfeld, R.
S. (1990). The effect of smoking on serum progesterone, estradiol, and luteinizing hormone
levels over a menstrual cycle in normal women. Steroids, 55(11), 507-511.
doi:10.1016/0039-128x(90)90089-t
19. Berta, L., Frairia, R., Fortunati, N., Fazzari, A., & Gaidano, G. (1992). Smoking effects on
the hormonal balance of fertile women. Hormone research, 37(1-2), 45-48.
doi:10.1159/000182280
20. Daniel, M., Martin, A. D., & Faiman, C. (1992). Sex hormones and adipose tissue
distribution in premenopausal cigarette smokers. International journal of obesity and
related metabolic disorders : journal of the International Association for the Study of
Obesity, 16(4), 245-254.
21. Ruiz, E., Osorio, E., & Ortega, E. (1992). Androgenic status in cyclic and postmenopausal
women: a comparison between smokers and nonsmokers. Biochemistry international,
27(5), 841-845.
22. Ortego-Centeno, N., Muñoz-Torres, M., Hernandez-Quero, J., Jurado-Duce, A., & de la
Higuera Torres-Puchol, J. (1994). Bone mineral density, sex steroids, and mineral
metabolism in premenopausal smokers. Calcified tissue international, 55(6), 403-407.
doi:10.1007/bf00298551
46
23. Key, T. J., Pike, M. C., Brown, J. B., Hermon, C., Allen, D. S., & Wang, D. Y. (1996).
Cigarette smoking and urinary oestrogen excretion in premenopausal and post-menopausal
women. British journal of cancer, 74(8), 1313-1316. doi:10.1038/bjc.1996.536
24. Westhoff, C., Gentile, G., Lee, J., Zacur, H., & Helbig, D. (1996). Predictors of ovarian
steroid secretion in reproductive-age women. American journal of epidemiology, 144(4),
381-388. doi:10.1093/oxfordjournals.aje.a008939
25. Alberti-Fidanza, A., Fruttini, D., & Servili, M. (1998). Gustatory and food habit changes
during the menstrual cycle. International journal for vitamin and nutrition research.
Internationale Zeitschrift fur Vitamin- und Ernahrungsforschung. Journal international de
vitaminologie et de nutrition, 68(2), 149-153.
26. Lucero, J., Harlow, B. L., Barbieri, R. L., Sluss, P., & Cramer, D. W. (2001). Early
follicular phase hormone levels in relation to patterns of alcohol, tobacco, and coffee use.
Fertility and sterility, 76(4), 723-729. doi:10.1016/s0015-0282(01)02005-2
27. Freour, T., Masson, D., Mirallie, S., Jean, M., Bach, K., Dejoie, T., & Barriere, P. (2008).
Active smoking compromises IVF outcome and affects ovarian reserve. Reproductive
biomedicine online, 16(1), 96-102. doi:10.1016/s1472-6483(10)60561-5
28. Dušková, M., Simůnková, K., Hill, M., Velíková, M., Kubátová, J., Kancheva, L., . . .
Pařízek, A. (2012). Chronic cigarette smoking alters circulating sex hormones and
neuroactive steroids in premenopausal women. Physiological research, 61(1), 97-111.
29. Gu, F., Caporaso, N. E., Schairer, C., Fortner, R. T., Xu, X., Hankinson, S. E., . . . Ziegler,
R. G. (2013). Urinary concentrations of estrogens and estrogen metabolites and smoking in
caucasian women. Cancer epidemiology, biomarkers & prevention : a publication of the
American Association for Cancer Research, cosponsored by the American Society of
Preventive Oncology, 22(1), 58-68. doi:10.1158/1055-9965.EPI-12-0909
30. Ellberg, C., Olsson, H., & Jernström, H. (2018). Current smoking is associated with a
larger waist circumference and a more androgenic profile in young healthy women from
high-risk breast cancer families. Cancer causes & control : CCC, 29(2), 243-251.
doi:10.1007/s10552-017-0999-3
31. McTiernan, A., Wu, L., Barnabei, V. M., Chen, C., Hendrix, S., Modugno, F., . . .
Investigators, W. H. I. (2008). Relation of demographic factors, menstrual history,
reproduction and medication use to sex hormone levels in postmenopausal women. Breast
Cancer Res Treat, 108(2), 217-231. doi:10.1007/s10549-007-9588-6
32. Kabat, G. C., Kim, M., Phipps, A. I., Li, C. I., Messina, C. R., Wactawski-Wende, J., . . .
Rohan, T. E. (2011). Smoking and alcohol consumption in relation to risk of triple-
negative breast cancer in a cohort of postmenopausal women. Cancer Causes Control,
22(5), 775-783. doi:10.1007/s10552-011-9750-7
47
33. Berstad, P., Ma, H., Bernstein, L., & Ursin, G. (2008). Alcohol intake and breast cancer
risk among young women. Breast Cancer Res Treat, 108(1), 113-120. doi:10.1007/s10549-
007-9578-8
34. Hamajima, N., Hirose, K., Tajima, K., Rohan, T., Calle, E. E., Heath, C. W., Jr., . . .
Collaborative Group on Hormonal Factors in Breast, C. (2002). Alcohol, tobacco and
breast cancer--collaborative reanalysis of individual data from 53 epidemiological studies,
including 58,515 women with breast cancer and 95,067 women without the disease. Br J
Cancer, 87(11), 1234-1245. doi:10.1038/sj.bjc.6600596
35. Hirko, K. A., Chen, W. Y., Willett, W. C., Rosner, B. A., Hankinson, S. E., Beck, A. H., . .
. Eliassen, A. H. (2016). Alcohol consumption and risk of breast cancer by molecular
subtype: Prospective analysis of the nurses' health study after 26 years of follow-up. Int J
Cancer, 138(5), 1094-1101. doi:10.1002/ijc.29861
36. Tworoger, S. S., Missmer, S. A., Eliassen, A. H., Barbieri, R. L., Dowsett, M., &
Hankinson, S. E. (2007). Physical activity and inactivity in relation to sex hormone,
prolactin, and insulin-like growth factor concentrations in premenopausal women -
exercise and premenopausal hormones. Cancer Causes Control, 18(7), 743-752.
doi:10.1007/s10552-007-9017-5
37. Tin Tin, S., Reeves, G. K., & Key, T. J. (2020). Body size and composition, physical
activity and sedentary time in relation to endogenous hormones in premenopausal and
postmenopausal women: Findings from the UK Biobank. Int J Cancer, 147(8), 2101-2115.
doi:10.1002/ijc.33010
38. Wetzels, J. J., Kremers, S. P., Vitoria, P. D., & de Vries, H. (2003). The alcohol-tobacco
relationship: a prospective study among adolescents in six European countries. Addiction,
98(12), 1755-1763. doi:10.1111/j.1360-0443.2003.00553.x
39. Sneve, M., & Jorde, R. (2008). Cross-sectional study on the relationship between body
mass index and smoking, and longitudinal changes in body mass index in relation to
change in smoking status: the Tromso Study. Scand J Public Health, 36(4), 397-407.
doi:10.1177/1403494807088453
40. Veldheer, S., Yingst, J., Zhu, J., & Foulds, J. (2015). Ten-year weight gain in smokers who
quit, smokers who continued smoking and never smokers in the United States, NHANES
2003-2012. Int J Obes (Lond), 39(12), 1727-1732. doi:10.1038/ijo.2015.127
41. Jackson, K. M., Sher, K. J., Cooper, M. L., & Wood, P. K. (2002). Adolescent alcohol and
tobacco use: onset, persistence and trajectories of use across two samples. Addiction, 97(5),
517-531. doi:10.1046/j.1360-0443.2002.00082.x
42. Brand, J. S., Chan, M. F., Dowsett, M., Folkerd, E., Wareham, N. J., Luben, R. N., . . .
Khaw, K. T. (2011). Cigarette smoking and endogenous sex hormones in postmenopausal
women. J Clin Endocrinol Metab, 96(10), 3184-3192. doi:10.1210/jc.2011-1165

48
Figure 2.1 Flowchart of selection process for articles included in assessment of smoking and sex
hormone level

a
Total number of analytes exceeds 19 as some articles provided measures for multiple sex hormones/proteins.

864 articles screened
461 duplicate articles
removed
19 articles included in the
meta-analysis

By sex hormone
a
:
16 Estradiol
9 Progesterone
9 Testosterone
2 DHEA
3 DHEAS
9 SHBG

64 articles excluded

19 Population-level characteristics
did not meet inclusion criteria
16 No report of serum/urinary
hormone levels for
smokers/nonsmokers
11 Oral contraceptive use not
excluded
6 Review article or intervention
5 Cannot contact author for study
information
5 Same population as another
article that was included
2 Article language not English

81 articles eligible for full
text review

2 articles identified
from reference list
reviews
783 articles deemed
irrelevant
1325 articles identified
from PubMed search

49
Table 2.1 Characteristics of studies selected for inclusion with measures of hormone levels by smoking status
Study Country
Biospecimen
sample type
Timing of
measure(s)
Age, years
a

nonsmokers / smokers
Covariates included
Number of
nonsmokers
Number of
smokers
Estradiol

Michnovicz, 1988 USA Urine Follicular 21 - 44 None 13 13
Longcope, 1988 USA Blood Varied 50.2 (0.6) / 49.4 (0.8) None 36 11
Daniel, 1992 Canada Blood Follicular 28.7 (5.2) / 29.5 (3.6) Sum of skinfolds 27 25
Berta, 1992 Italy Blood Luteal 33.0 (3.8) / 32.8 (3.5) None 447 237
Ortego-Centeno, 1994 Spain Blood Follicular 28.3 (7.6) / 28.2 (6.2) None 15 16
Key, 1996 UK Urine Follicular 42.4 (4.0) / 41.5 (4.8) Age and BMI 46 29
Key, 1996 UK Urine Luteal 42.4 (4.0) / 41.5 (4.8) Age and BMI 68 24
Westoff, 1996 USA Urine Luteal 21 - 36 Urinary creatinine levels 122 48
Alberti-Fidanza, 1998 Italy Blood Follicular 25-37 / 23-31 None 2 6
Alberti-Fidanza, 1998 Italy Blood Luteal 25-37 / 23-31 None 2 6
Lucero, 2001 USA Blood Follicular 36 - 45 Age 274 52
Freour, 2008 France Blood Follicular 31.4 (4.5) / 31.8 (5.1) None 71 40
Dafopoulos, 2010 Greece Blood Follicular 20 - 49 None 78 59
Soldin, 2011 USA Blood Follicular 18 - 45 None 100 107
Duskova, 2012 Czech Republic Blood Follicular
39.5 (31.8, 44.1) / 39.7
(32.2, 48.1)
Age (matched) 10 10
Duskova, 2012 Czech Republic Blood Luteal
35.3 (31.7, 44.1) / 38.7
(35, 43.5)
Age (matched) 10 8
EHBCCG, 2013 International Blood Varied 35.6 - 42.2
Study, age, BMI and
phase of cycle
409 1555
Gu, 2013 USA Urine Luteal 42.7 (3.9) / 42.9 (2.9)
Age, BMI, alcohol use,
physical activity, first
morning urine, and luteal
day at urine collection
428 35
Ellberg, 2018 Sweden Blood Varied 29 (23, 34) / 29 (25, 35)
Adjusted for age,
nulliparity, weight,
height, WHR, and days
until the next menstrual
period
110 32
Progesterone

Zumoff, 1990 USA Blood Follicular 20 - 35 None 8 8
Berta, 1992 Italy Blood Luteal 33.0 (3.8) / 32.8 (3.5) None 447 237
Ortego-Centeno, 1994 Spain Blood Follicular 28.3 (7.6) / 28.2 (6.2) None 15 16
Westoff, 1996 USA Blood Luteal 21 - 36 None 122 48
Alberti-Fidanza, 1998 Italy Blood Follicular 25-37 / 23-31 None 2 6

50
Study Country
Biospecimen
sample type
Timing of
measure(s)
Age, years
a

nonsmokers / smokers
Covariates included
Number of
nonsmokers
Number of
smokers
Alberti-Fidanza, 1998 Italy Blood Luteal 25-37 / 23-31 None 2 6
Dafopoulos, 2010 Greece Blood Follicular 20 - 49 None 78 59
Soldin, 2011 USA Blood Follicular 18 - 45 None 100 107
Duskova, 2012 Czech Republic Blood Follicular
39.5 (31.8, 44.1) / 39.7
(32.2, 48.1)
Age (matched) 10 10
Duskova, 2012 Czech Republic Blood Luteal
35.3 (31.7, 44.1) / 38.7
(35, 43.5)
Age (matched) 10 8
EHBCCG, 2013 International Blood Varied 35.6 - 42.2
Study, age, BMI and
phase of cycle
296 1293
Testosterone

Longcope, 1988 USA Blood Varied 50.2 (0.6) / 49.4 (0.8) None 36 11
Daniel, 1992 Canada Blood Follicular 28.7 (5.2) / 29.5 (3.6) Sum of skinfolds 27 25
Ruiz, 1992 Spain Blood Varied NR None 27 15
Ortego-Centeno, 1994 Spain Blood Follicular 28.3 (7.6) / 28.2 (6.2) None 15 16
Barbieri, 2005 USA Blood Varied <30 - >39 None 257 162
Soldin, 2011 USA Blood Follicular 18 - 45 None 100 107
Duskova, 2012 Czech Republic Blood Follicular
39.5 (31.8, 44.1) / 39.7
(32.2, 48.1)
Age (matched) 10 10
Duskova, 2012 Czech Republic Blood Luteal
35.3 (31.7, 44.1) / 38.7
(35, 43.5)
Age (matched) 10 8
EHBCCG, 2013 International Blood Varied 35.6 - 42.2
Study, age, BMI and
phase of cycle
507 1931
Ellberg, 2018 Sweden Blood Varied 29 (23, 34) / 29 (25, 35)
Adjusted for age,
nulliparity, weight,
height, WHR, and days
until the next menstrual
period
110 32
DHEA

Soldin, 2011 USA Blood Follicular 18 - 45 None 100 107
Duskova, 2012 Czech Republic Blood Follicular
39.5 (31.8, 44.1) / 39.7
(32.2, 48.1)
Age (matched) 10 10
Duskova, 2012 Czech Republic Blood Luteal
35.3 (31.7, 44.1) / 38.7
(35, 43.5)
Age (matched) 10 8
DHEAS

Ortego-Centeno, 1994 Spain Blood Follicular 28.3 (7.6) / 28.2 (6.2) None 15 16
Soldin, 2011 USA Blood Follicular 18 - 45 None 100 107
EHBCCG, 2013 International Blood Varied 35.6 - 42.2
Study, age, BMI and
phase of cycle
348 1537

51
Study Country
Biospecimen
sample type
Timing of
measure(s)
Age, years
a

nonsmokers / smokers
Covariates included
Number of
nonsmokers
Number of
smokers
SHBG

Daniel, 1992 Canada Blood Follicular 28.7 (5.2) / 29.5 (3.6) Sum of skinfolds 27 25
Berta, 1992 Italy Blood Luteal 33.0 (3.8) / 32.8 (3.5) None 447 237
Ruiz, 1992 Spain Blood Varied NR None 27 15
Ortego-Centeno, 1994 Spain Blood Follicular 28.3 (7.6) / 28.2 (6.2) None 15 16
Lucero, 2001 USA Blood Follicular 36 - 45 Age 274 52
Barbieri, 2005 USA Blood Varied <30 - >39 None 257 162
Duskova, 2012 Czech Republic Blood Follicular
39.5 (31.8, 44.1) / 39.7
(32.2, 48.1)
Age (matched) 10 10
Duskova, 2012 Czech Republic Blood Luteal
35.3 (31.7, 44.1) / 38.7
(35, 43.5)
Age (matched) 10 8
EHBCCG, 2013 International Blood Varied 35.6 - 42.2
Study, age, BMI and
phase of cycle
517 1965
Ellberg, 2018 Sweden Blood Varied 29 (23, 34) / 29 (25, 35)
Adjusted for age,
nulliparity, weight,
height, WHR, and days
until the next menstrual
period
110 32
a
Age is presented as mean (SD), median (IQR), or range.
Abbreviations: BMI, body mass index; DHEA, dehydroepiandrosterone; DHEAS, dehydroepiandrosterone sulfate; EHBCCG, Endogenous Hormones and Breast Cancer
Collaborative Group; IQR, interquartile range; NR, not reported; SD, standard deviation; SHBG, sex hormone-binding globulin; UK, United Kingdom; USA, United States of
America; WHR, waist-to-hip ratio.

52

Table 2.2. Summary of the standardized mean differences between smokers and nonsmokers for
random effects among studies selected for meta-analysis by hormone
k SMD (95% CI)
a
I
2
(%) Phet
Follicular
Estradiol 10 0.10 (-0.17, 0.38) 67.6 0.001
Progesterone 6 0.27 (-0.17, 0.72) 67.3 0.01
Testosterone 4 -0.03 (-0.48, 0.43) 61.1 0.05
DHEA 2 0.23 (-0.71, 1.18) 75.6 0.04
DHEAS 2 -0.05 (-0.30, 0.20) 0.0 0.98
SHBG 4 0.15 (-0.35, 0.65) 66.5 0.03
Varied
Estradiol 3 -0.07 (-0.18, 0.03) 0.0 0.52
Progesterone 0 --- --- ---
Testosterone 5 0.14 (0.0005, 0.29) 31.4 0.21
DHEA 0 --- --- ---
DHEAS 1 0.15 (0.04, 0.27) --- ---
SHBG 4 -0.07 (-0.25, 0.11) 51.3 0.10
Luteal
Estradiol 6 -0.17 (-0.53, 0.19) 76.1 0.001
Progesterone 5 0.02 (-0.28, 0.32) 79.6 0.001
Testosterone 1 0.91 (-0.07, 1.89) --- ---
DHEA 1 1.90 (0.76, 3.04) --- ---
DHEAS 0 --- --- ---
SHBG 2 -0.19 (-0.91, 0.53) 60.5 0.11
Overall
Estradiol 19 -0.03 (-0.18, 0.13) 65.5 <0.001
Progesterone 11 0.12 (-0.12, 0.36) 75.0 <0.001
Testosterone 10 0.14 (-0.03, 0.30) 48.9 0.04
DHEA 3 0.76 (-0.45, 1.97) 86.8 0.001
DHEAS 3 0.12 (0.01, 0.22) 0.0 0.37
SHBG 10 -0.01 (-0.15, 0.13) 52.4 0.03
a
Bold values note significant differences in standardized mean hormone level by smoking status at P<0.05.
Abbreviations: CI, confidence interval; DHEA, dehydroepiandrosterone; DHEAS, dehydroepiandrosterone sulfate; het,
heterogeneity; n, number of studies; SHBG, sex hormone-binding globulin; SMD, standardized mean difference.

53

Figure 2.2. Forest plots depicting the standardized mean differences between smokers and
nonsmokers for random effects among studies selected for meta-analysis by hormone
a. Estradiol b. Progesterone

c. Testosterone d. DHEA

e. DHEAS f. SHBG

54

Abbreviations: CI, confidence interval; DHEA, dehydroepiandrosterone; DHEAS, dehydroepiandrosterone sulfate; EHBCCG,
Endogenous Hormones and Breast Cancer Collaborative Group; SHBG, sex hormone-binding globulin; SMD, standardized mean
difference.

55

Figure 2.3. Funnel plots depicting bias in the standardized mean differences by menstrual phase
and hormone
a. Estradiol b. Progesterone

c. Testosterone d. DHEA

e. DHEAS f. SHBG

56

Abbreviations CI, confidence interval; DHEA, dehydroepiandrosterone; DHEAS, dehydroepiandrosterone sulfate; SHBG, sex
hormone-binding globulin; SMD, standardized mean difference.

57
Figure 2.4. Power curves for estradiol by degree of heterogeneity and menstrual phase
a. Overall

b. Follicular

c. Varied

d. Luteal

58
Supplementary Table 2.1. PubMed search terms and results
Search term
Last search: December 1, 2019
Number
of articles
retrieved
(Female[Mesh] AND "Gonadal Steroid Hormones"[Mesh] AND ("Tobacco
Smoking"[Mesh] OR "Cigarette Smoking"[Mesh] OR "Smoking"[Mesh])) NOT
"Mice"[Mesh]
547
(Female[Mesh] AND "Estrogens"[Mesh] AND ("Tobacco Smoking"[Mesh] OR
"Cigarette Smoking"[Mesh] OR "Smoking"[Mesh])) NOT "Mice"[Mesh]
363
(Female[Mesh] AND "Progesterone"[Mesh] AND ("Tobacco Smoking"[Mesh]
OR "Cigarette Smoking"[Mesh] OR "Smoking"[Mesh])) NOT "Mice"[Mesh]
131
(Female[Mesh] AND "Androgens"[Mesh] AND ("Tobacco Smoking"[Mesh] OR
"Cigarette Smoking"[Mesh] OR "Smoking"[Mesh])) NOT "Mice"[Mesh]
44
(Female[Mesh] AND "Testosterone"[Mesh] AND ("Tobacco Smoking"[Mesh]
OR "Cigarette Smoking"[Mesh] OR "Smoking"[Mesh])) NOT "Mice"[Mesh]
97
(Female[Mesh] AND "Dehydroepiandrosterone"[Mesh] AND ("Tobacco
Smoking"[Mesh] OR "Cigarette Smoking"[Mesh] OR "Smoking"[Mesh])) NOT
"Mice"[Mesh]
60
(Female[Mesh] AND "Dehydroepiandrosterone Sulfate"[Mesh] AND ("Tobacco
Smoking"[Mesh] OR "Cigarette Smoking"[Mesh] OR "Smoking"[Mesh])) NOT
"Mice"[Mesh]
39
(Female[Mesh] AND "Sex Hormone-Binding Globulin"[Mesh] AND ("Tobacco
Smoking"[Mesh] OR "Cigarette Smoking"[Mesh] OR "Smoking"[Mesh])) NOT
"Mice"[Mesh]
44

59
Supplementary Table 2.2. Equations for conversion from Wan et.al., 2014
Conversion Equation
Mean from median and IQR 𝑋 ̅
≈
𝑞 1
+ 𝑚 + 𝑞 3
3

SD from median and IQR
𝑆 ≈
𝑞 3
− 𝑞 1
2Φ
−1
(
0.75𝑛 − 0.125
𝑛 + 0.25
)

Mean from median and range
𝑋 ̅
≈
𝑎 + 2𝑚 + 𝑏 4

SD from median and range
𝑆 ≈
𝑏 − 𝑎 2Φ
−1
(
𝑛 − 0.375
𝑛 + 0.25
)

Abbreviations: a, minimum value; b, maximum value; IQR, interquartile range; m, median; n, sample size; q1, first quartile; q3,
third quartile; SD, standard deviation.

60
Supplementary Table 2.3. Summary of the standardized mean differences between smokers and
nonsmokers for random effects among studies with covariate-adjusted measures selected for
meta-analysis by hormone
k SMD (95% CI)
a
I
2
(%) Phet
Follicular
Estradiol 4 -0.17 (-0.74, 0.40) 80.1 0.002
Progesterone 1 1.31 (0.34, 2.29) --- ---
Testosterone 2 0.27 (-1.15, 1.69) 85.3 0.009
DHEA 1 0.83 (-0.09, 1.75) --- ---
DHEAS 0 --- --- --
SHBG 3 0.01 (-0.56, 0.57) 68.6 0.04
Varied
Estradiol 2 -0.08 (-0.18, 0.03) 0.0 0.43
Progesterone 0 --- --- ---
Testosterone 2 0.16 (-0.12, 0.45) 56.4 0.13
DHEA 0 --- --- ---
DHEAS 1 0.15 (0.04, 0.27) --- ---
SHBG 2 0.03 (-0.07, 0.12) 0.0 0.76
Luteal
Estradiol 4 -0.26 (-0.53, 0.01) 32.7 0.22
Progesterone 2 0.39 (-0.60, 1.38) 75.7 0.04
Testosterone 1 0.91 (-0.07, 1.89) --- ---
DHEA 1 1.90 (0.76, 3.04) --- ---
DHEAS 0 --- --- ---
SHBG 1 -0.74 (-1.70, 0.23) --- ---
Overall
Estradiol 10 -0.18 (-0.37, 0.01) 60.0 0.007
Progesterone 3 0.68 (-0.28, 1.64) 81.6 0.004
Testosterone 5 0.24 (-0.12, 0.60) 66.8 0.02
DHEA 2 1.31 (0.27, 2.35) 51.3 0.15
DHEAS 1 0.03 (-0.20, 0.26) --- ---
SHBG 6 0.004 (-0.20, 0.21) 43.0 0.12
a
Bold values note significant differences in standardized mean hormone level by smoking status at P<0.05.
Abbreviations: CI, confidence interval; DHEA, dehydroepiandrosterone; DHEAS, dehydroepiandrosterone sulfate; het,
heterogeneity; n, number of studies; SHBG, sex hormone-binding globulin; SMD, standardized mean difference.

61
Supplementary Figure 2.1. Forest plots depicting the standardized mean differences between
smokers and nonsmokers for random effects among studies with covariate-adjusted measures
selected for meta-analysis by hormone
a. Estradiol b. Progesterone

c. Testosterone d. DHEA

e. DHEAS f. SHBG

62

Abbreviations CI, confidence interval; DHEA, dehydroepiandrosterone; DHEAS, dehydroepiandrosterone sulfate; EHBCCG,
Endogenous Hormones and Breast Cancer Collaborative Group; SHBG, sex hormone-binding globulin; SMD, standardized mean
difference.

63
CHAPTER THREE: Using the Young Women’s Health History Study to Investigate the
Role of Tobacco Exposure and Breast Cancer Risk Among Young Women

3.1 Study overview
The Young Women’s Health History Study (YWHHS) is a completed population-based
case-control study of invasive breast cancer among young non-Hispanic (NH) Black and NH
White women under age 50 years residing in the Los Angeles County (LA County) and
Metropolitan Detroit Surveillance, Epidemiology, and End Results (SEER) Registry areas. In
total, 1,812 women with breast cancer diagnosed from 2010 to 2015 were enrolled (1,130 NH
White and 682 NH Black), as were a population-based sample of 1,381 control women (716 NH
White and 665 NH Black) frequency-matched to cases on age, race, and SEER registry area. Dr.
Ellen Velie from the University of Wisconsin, Milwaukee was the overall Principal Investigator
(PI) of the YWHHS, and Dr. Ann Hamilton was a co-investigator and the PI of the LA County
subcontract (R01CA136861, Ellen Velie, PhD, PI). The fieldwork for the study was conducted
between 2010-2016 and data are now available for analysis. The objective of this dissertation
work is to conduct a thorough investigation on the association between tobacco exposure and
breast cancer risk using data from the YWHHS.

3.2 Case / control definition and eligibility criteria
Cases in the LA County and Metropolitan Detroit SEER Registries were identified by
rapid case ascertainment. Cases diagnosed with histologically confirmed invasive primary breast
cancer between September 1, 2010 and August 31, 2015, with no diagnosis of invasive or in situ
breast cancer prior to the reference date (date of first microscopic cytologic/histologic breast

64
cancer diagnosis) and no prior diagnosis of other invasive cancers, except cervical in situ or
common skin cancer, prior to the reference date were eligible for inclusion. All NH Black
women 20-49 years of age, all NH White women 20-45 years of age and a sample of NH White
women 45-49 years of age were included. Additional eligibility criteria included: living cases,
born in the United States (US), resident of a study region (LA County or Metropolitan Detroit)
on date of diagnosis, resident of a household (i.e., not institutionalized), able to be interviewed in
English, physically and mentally capable of completing the interview, and having access to a
telephone to be contacted by study personnel.
An area-based control sampling design was developed to identify women without any
history of invasive or in situ breast cancer diagnosis (controls) who otherwise met the same
eligibility criteria as cases and were frequency matched to cases on race, region, and five-year
age group. Controls in each region were recruited by the Westat Research Corporation (Westat).
A multi-stage area probability sampling method was developed to identify a sample of controls
representative of NH White and NH Black women in LA County and Metropolitan Detroit.
Westat conducted rosters of all households, as well as screening interviews to identify eligible
participants. Eligible control participants were provided to the PI and contact information was
uploaded to a central tracking system for interviewers in each study region to follow up with
potential participants. The study reference date for each control was set as the date on which
Westat conducted the in-person household screening interview minus four months.

3.3 Data collection
The study questionnaire was administered via Computer Assisted Personal Interview
(CAPI) and recorded on a laptop computer for both cases and controls. A life-history calendar

65
aided in the conduct of the interview and contained major life events to assist respondents in
remembering earlier life events. The interviewers collected demographic data, medical histories,
reproductive histories and information on lifestyle factors including physical activity, alcohol
consumption and tobacco use. Questionnaire items were identified from recent studies with
validated measures including the Women’s Contraceptive and Reproductive Experience
(Women's CARE) Study, the Long Island Breast Cancer Study Project, the National
Longitudinal Study of Adolescent Health, the Sister Study, and the Women’s Interview Study of
Health (1-5). The PhenX Toolkit, a resource for data collection protocols and measures of
phenotypes and exposures in biomedical research, was also used to inform the development of
the tobacco history section (6).
While administering questionnaires, study interviewers asked the participant if she would
be willing to ask her primary childhood caregiver (usually her mother) to complete a short,
mailed survey. The Caregiver’s Survey asked about the mother’s pregnancy history, prenatal
exposures and characteristics of the pregnancy with the participant, the participant’s childhood
medical history and lifestyle factors, and other family and sociodemographic characteristics.

3.4 Selection of YWHHS for dissertation research
The YWHHS was designed to evaluate multiple risk factors for breast cancer among a
population of young women. The study recruited a large, socioeconomically diverse sample of
NH Black and NH White women which provides a unique opportunity to evaluate racial and
socioeconomic disparities in breast cancer risk. A detailed risk factor questionnaire was
administered to obtain a thorough history of tobacco exposure including an assessment of self-
reported personal tobacco smoking (Appendix A1.1) prenatal smoke exposures (Appendix

66
A1.2). Tobacco exposures collected in the questionnaire included personal smoking history,
secondhand smoke exposures in childhood and adulthood, and prenatal smoking exposure. If the
participant’s primary childhood caregiver completed the Caregiver’s Survey, they also reported
on prenatal smoking exposure (Appendix A1.2). The availability of the caregiver’s information
made it possible to confirm the participant’s recall of childhood and prenatal smoke exposures
and to assess the reliability of the prenatal tobacco exposure responses.

3.5 Dissertation chapters based on YWHHS: An overview
Using the YWHHS data, two dissertation chapters were developed to evaluate the
association between lifetime tobacco exposure and early onset breast cancer risk accounting for
established breast cancer risk factors among Black and White women under 50 years old. We
evaluated breast cancer risk overall and by tumor subtype. Differences were assessed by race and
socioeconomic position (SEP).
Specifically, Chapter 4 will evaluate the association between lifetime personal tobacco
use and breast cancer risk among young women. Various measures of personal smoking history
prior to breast cancer diagnosis will be derived. Chapter 5 will evaluate the association between
two smoking exposures and breast cancer risk across the participants’ lifetime: lifetime
secondhand smoke exposure (including exposures in childhood and adulthood) and a composite
smoke exposure measure that will include lifetime personal and secondhand smoke exposures.
Lifetime smoking exposures will also be evaluated with an assessment for effect modification by
prenatal smoke exposures status. The details of the derivation of these tobacco exposure
measures are described in the following section.

67
3.6 Variable definitions
3.6.1 Clinical outcomes
As confirmed with SEER Registry data:
• Breast cancer status (categorical): control, case
• Breast cancer subtype (categorical): control, Luminal A, Luminal B, Human Epidermal
Growth Factor Receptor 2 (HER2)-type, and Triple Negative Breast Cancer (TNBC)

3.6.2 Tobacco exposures
Personal smoking history will be described with consideration of smoking status, smoking
intensity, and the timing of smoking initiation.
• Ever smoker status (categorical): never smoker, ever smoker
o Ever smokers described as women who have smoked at least one cigarette a day
for 6 months or more.
• Personal smoking status (categorical): never smoker, former smoker, current smoker
• Smoking intensity in pack-years (categorical): never smoker, <5, 5-19, ≥20
o Calculated as follows:
𝑝𝑎𝑐𝑘 -𝑦𝑒𝑎𝑟𝑠 =
# 𝑐𝑖𝑔𝑎𝑟𝑒𝑡𝑡𝑒𝑠 𝑠𝑚𝑜𝑘𝑒𝑑 𝑝𝑒𝑟 𝑑𝑎𝑦 20 𝑐𝑖𝑔𝑎𝑟𝑒𝑡𝑡𝑒𝑠 𝑝𝑒𝑟 𝑝𝑎𝑐𝑘 × # 𝑦𝑒𝑎𝑟𝑠 𝑜𝑓 𝑠𝑚𝑜𝑘𝑖𝑛𝑔
• Age smoking initiated in years (categorical): never smoker, <18, 18-24, ≥25
• Time since initiation in years (categorical): never smoker, <20, 20-29, ≥30
o Calculated as follows:
𝑦𝑒𝑎𝑟𝑠 = 𝑎𝑔𝑒 𝑎𝑡 𝑟𝑒𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑑𝑎𝑡𝑒 − 𝑎𝑔𝑒 𝑎𝑡 𝑠𝑚𝑜𝑘𝑖𝑛𝑔 𝑖𝑛𝑡𝑖𝑎𝑡𝑖𝑜𝑛
• Average number of cigarettes per day (categorical): never smoker, <5, 5-19, ≥20

68
• Smoking initiation timing – first full-term pregnancy (categorical): never smoker, after
first full-term pregnancy; before first full-term pregnancy
o Will be estimated among parous women only.
o Will be derived from data provided on age at first full-term pregnancy and age at
smoking initiation.

Prenatal tobacco exposure will be categorized as a binary yes/no variable as reported from the
participant questionnaires.

Secondhand smoke exposure will be derived using information from childhood and adulthood as
follows:
• Lifetime secondhand smoke exposure status (binary): no, yes.
o Duration of lifetime secondhand smoke exposure (categorical): No secondhand
exposure, <15 years, ≥15 years
▪ Cut point identified from the median duration of exposure among controls.
• Secondhand smoke exposure status in childhood (binary): no, yes.
o Duration of secondhand smoke exposure in childhood (categorical): No
secondhand exposure, <15 years, ≥15 years
▪ Cut point identified from the median duration of exposure among controls.
• Secondhand smoke exposure status in adulthood (binary): no, yes.
o Duration of secondhand smoke exposure in adulthood (categorical): No
secondhand exposure, <10 years, ≥10 years
▪ Cut point identified from the median duration of exposure among controls.

69

The composite variable for lifetime smoking exposures is adapted from methods used in the
Women’s Health Initiative Observational Study (7-10). In models to consider associations
between each of the smoking exposures and breast cancer risk, personal smoking history,
secondhand smoke exposure in childhood and secondhand smoke exposure in adulthood are
captured as binary yes/no variables.
• Lifetime composite smoke exposure (categorical): no secondhand exposure, childhood
exposure only, adulthood exposure only, childhood and adulthood exposure, ever smoker.
o The categories are mutually exclusive.
o The combinations of secondhand smoke exposures are developed among non-
smokers; ever smokers are captured in one category regardless of secondhand
smoke exposures.
o A sensitivity analysis of the association will be conducted to explore effect
modification by prenatal smoke exposure status.

A p-value for trend for each main exposure variable with an ordinal data structure will be based
on the Wald’s test for the variable in a continuous model including: smoking intensity, age at
smoking initiation, time since initiation, and average number of cigarettes smoked per day. The
list of questionnaire items that were used to derive the tobacco exposure variables is provided in
Appendix 1.

3.6.3 Matching characteristics and model covariates

70
For the study design, participants were frequency matched in 5-year intervals of age (20-
24, 25-29, 30-34, 35-39, 40-44, 45-49). Participants were also matched by region (Los Angeles
County and Metropolitan Detroit) and self-reported race (NH Black and NH White).
The covariates to be considered for population summaries and/or inclusion in
multivariable adjusted models are defined as follows:
• Household poverty level (binary): <200%, ≥200% of the federal poverty level (FPL)
o Derived from self-reported gross income and number of members in household to
calculate the percent of the FPL.
o Low SEP is defined as living <200% of the FPL (compared to living ≥200% of
the FPL)
• Education (categorical): high school or less, vocational school/associate’s degree/some
college, bachelor’s degree/some post-graduate, and master’s degree or higher.
• First-degree family history of breast cancer (categorical): no, yes, unknown
• Age at menarche in years (categorical): ≤11, 12, 13, ≥14
• Joint parity and age (in years) at first full-term birth (categorical): nulliparous, 1-2, <25;
1-2, ≥25; 3+, <25; 3+, ≥25
• Premenopausal status (binary): premenopausal, peri- & post menopausal
• Alcohol use in grams per day (categorical): 0 (Abstainers); 0.1-6.9, 7-13.9, 14-27.9, ≥28
• Body mass index (in kg/m
2
) 12 months before reference date (categorial): underweight
(<18.5), normal (18.5-24.9), overweight (25-29.9), obese (≥30).

3.6.4 Strengths and limitations of the measures

71
There are some limitations in the available data. Quantitative information on the
frequency or timing (trimester) of prenatal smoke exposure (i.e., mother’s smoking) is not
available. In addition, we only asked about cigarette smoking and did not have information on
other tobacco product use/exposures (e.g., cigars, pipes, chewing tobacco and snuff). There is a
potential for differential misclassification of intensity of tobacco exposure if breast cancer cases
who are smokers are more likely to be polytobacco users. Results from the 2017 National Survey
on Drug Use and Health show that 57.2% of women ages 18 years and older had used cigarettes
in their lifetime and 18.3% of women ages 18 years and older had used cigars (11). In another
assessment of trends in use of little cigars and cigarillos, about 70% of female adolescent and
young adult cigar smokers also smoked cigarettes (12). Furthermore, data from the Population
Assessment of Tobacco and Health (PATH) Study revealed that 61% of the sample used
cigarettes only and use of cigars, cigarillos, filtered cigar, and pipes was positively correlated
with the use of another combustible tobacco product (potentially cigarettes) (13). Therefore, the
analysis on cigarette use will capture a substantial degree of the tobacco exposure in this
population with some potential for underestimating the association.
There may be some unmeasured confounding factors. Some known risk factors to breast
cancer were not obtained in this study, including childhood chest radiation, history of benign
breast disease and specific genetic factors (BRCA1/2, TP53 mutations). In addition, information
was not available on health care access and utilization that could contribute to disparities in
disease risk (14; 15). Information on oral contraceptive use and history of breastfeeding were not
available for analysis from the data source.
For prenatal smoke exposures, about 60% of the primary childhood caregivers of subjects
participated but the proportion varied by subgroup with the highest response rate among

72
childhood caregivers of White cases (73%) and the lowest response rate among caregivers of
Black controls (25%). We plan to minimize bias from missing responses by using the
participant’s self-report. According to a surveillance report using data from the Pregnancy Risk
Assessment Monitoring System (PRAMS) questionnaire, 54.3% of women across all
racial/ethnic groups who smoked during their pregnancy quit by the last trimester and 12.3% of
women (all races) reported smoking during the last 3 months of pregnancy (16). Therefore, we
will assess the validity of using the participants’ report by evaluating the concordance in the
primary childhood caregiver and participant’s account of prenatal smoke exposure among
records with both surveys completed.
Risk estimates may be unstable because of small samples in some of the stratified
analyses defined by race/SEP subpopulations. Although we anticipate a lack of sufficient power
to statistically detect an association, we believe these findings are important and will suggest a
direction for future analyses.

3.7 Summary
The YWHHS is an ideal resource for evaluating the role of tobacco exposure, among
other risk factors, that may contribute to racial and socioeconomic disparities in risk of young-
onset breast cancer. The information collected from the YWHHS study interviews will be used
to derive measures of prenatal smoke exposure, lifetime secondhand smoke exposure in
childhood and adulthood, and personal tobacco use, including timing and intensity of use. This
dissertation explores the association between tobacco exposures at different time periods in life,
including a characterization of composite lifetime smoke exposure, and risk of breast cancer
among young women by race and SEP. The findings of this study will represent new knowledge

73
on the timing and types of tobacco exposures that are associated with young-onset breast cancer
and whether subpopulations may be particularly vulnerable to these exposures.

74
3.8 References
1. Marchbanks, P. A., McDonald, J. A., Wilson, H. G., Burnett, N. M., Daling, J. R.,
Bernstein, L., . . . Spirtas, R. (2002). The NICHD Women's Contraceptive and
Reproductive Experiences Study: methods and operational results. Ann Epidemiol, 12(4),
213-221.
2. Gammon, M. D., Neugut, A. I., Santella, R. M., Teitelbaum, S. L., Britton, J. A., Terry, M.
B., . . . Obrams, G. I. (2002). The Long Island Breast Cancer Study Project: description of
a multi-institutional collaboration to identify environmental risk factors for breast cancer.
Breast Cancer Res Treat, 74(3), 235-254.
3. Resnick, M. D., Bearman, P. S., Blum, R. W., Bauman, K. E., Harris, K. M., Jones, J., . . .
Udry, J. R. (1997). Protecting adolescents from harm. Findings from the National
Longitudinal Study on Adolescent Health. JAMA, 278(10), 823-832.
4. Brinton, L. A., Daling, J. R., Liff, J. M., Schoenberg, J. B., Malone, K. E., Stanford, J. L., .
. . Hoover, R. N. (1995). Oral contraceptives and breast cancer risk among younger
women. J Natl Cancer Inst, 87(11), 827-835.
5. Sandler, D. P., Hodgson, M. E., Deming-Halverson, S. L., Juras, P. S., D'Aloisio, A. A.,
Suarez, L. M., . . . Sister Study Research, T. (2017). The Sister Study Cohort: Baseline
Methods and Participant Characteristics. Environ Health Perspect, 125(12), 127003.
doi:10.1289/EHP1923
6. Hamilton, C. M., Strader, L. C., Pratt, J. G., Maiese, D., Hendershot, T., Kwok, R. K., . . .
Haines, J. (2011). The PhenX Toolkit: get the most from your measures. Am J Epidemiol,
174(3), 253-260. doi:10.1093/aje/kwr193
7. Hyland, A., Piazza, K., Hovey, K. M., Tindle, H. A., Manson, J. E., Messina, C., . . .
Wactawski-Wende, J. (2016). Associations between lifetime tobacco exposure with
infertility and age at natural menopause: the Women's Health Initiative Observational
Study. Tob Control, 25(6), 706-714. doi:10.1136/tobaccocontrol-2015-052510
8. Hyland, A., Piazza, K. M., Hovey, K. M., Ockene, J. K., Andrews, C. A., Rivard, C., &
Wactawski-Wende, J. (2015). Associations of lifetime active and passive smoking with
spontaneous abortion, stillbirth and tubal ectopic pregnancy: a cross-sectional analysis of
historical data from the Women's Health Initiative. Tob Control, 24(4), 328-335.
doi:10.1136/tobaccocontrol-2013-051458
9. Piazza, K. M., Wactawski-Wende, J., DeBon, M. W., Hovey, K. M., Rivard, C. L., Smith,
D. M., & Hyland, A. J. (2016). Inhaled medication usage in post-menopausal women and
lifetime tobacco smoke exposure: The Women's Health Initiative Observational Study.
Maturitas, 90, 42-48. doi:10.1016/j.maturitas.2016.05.008
10. Wang, A., Kubo, J., Luo, J., Desai, M., Hedlin, H., Henderson, M., . . . Wakelee, H. A.
(2015). Active and passive smoking in relation to lung cancer incidence in the Women's

75
Health Initiative Observational Study prospective cohort. Ann Oncol, 26(1), 221-230.
doi:10.1093/annonc/mdu470
11. Center for Behavioral Health Statistics and Quality. (2018). 2017 National Survey on Drug
Use and Health: Detailed Tables. Retrieved from Rockville, MD:
https://www.samhsa.gov/data/sites/default/files/cbhsq-
reports/NSDUHDetailedTabs2017/NSDUHDetailedTabs2017.pdf
12. Messer, K., White, M. M., Strong, D. R., Wang, B., Shi, Y., Conway, K. P., & Pierce, J. P.
(2015). Trends in use of little cigars or cigarillos and cigarettes among U.S. smokers, 2002-
2011. Nicotine Tob Res, 17(5), 515-523. doi:10.1093/ntr/ntu179
13. Kypriotakis, G., Robinson, J. D., Green, C. E., & Cinciripini, P. M. (2018). Patterns of
Tobacco Product Use and Correlates Among Adults in the Population Assessment of
Tobacco and Health (PATH) Study: A Latent Class Analysis. Nicotine Tob Res,
20(suppl_1), S81-S87. doi:10.1093/ntr/nty025
14. DeSantis, C. E., Fedewa, S. A., Goding Sauer, A., Kramer, J. L., Smith, R. A., & Jemal, A.
(2016). Breast cancer statistics, 2015: Convergence of incidence rates between black and
white women. CA Cancer J Clin, 66(1), 31-42. doi:10.3322/caac.21320
15. Sheppard, V. B., Mays, D., LaVeist, T., & Tercyak, K. P. (2013). Medical mistrust
influences black women's level of engagement in BRCA 1/2 genetic counseling and
testing. J Natl Med Assoc, 105(1), 17-22.
16. Tong, V. T., Dietz, P. M., Morrow, B., D'Angelo, D. V., Farr, S. L., Rockhill, K. M., . . .
Centers for Disease, C. P. (2013). Trends in smoking before, during, and after pregnancy--
Pregnancy Risk Assessment Monitoring System, United States, 40 sites, 2000-2010.
MMWR Surveill Summ, 62(6), 1-19.

76
CHAPTER FOUR: Lifetime cigarette smoking and risk for breast cancer subtypes among
Non-Hispanic Black and White women in the Young Women’s Health History Study

Authors: Ugonna Ihenacho
1
, Ann S. Hamilton
1
, Wendy J. Mack
1
, Anna H. Wu
1
, Jennifer B.
Unger
1
, Dorothy R. Pathak
2
, Richard T. Houang
3
, Michael F. Press
4
, Kendra L. Schwartz
5
, Lydia
R. Marcus
6
, Ellen M. Velie
6,7*

Institutional affiliations:
1. Department of Preventive Medicine, Keck School of Medicine, University of Southern
California, Los Angeles, CA, USA
2. Department of Epidemiology and Biostatistics, College of Human Medicine, Michigan State
University, East Lansing, MI, USA
3. Center for the Study of Curriculum, College of Education, Michigan State University, East
Lansing, MI, USA
4. Department of Pathology, Keck School of Medicine, University of Southern California, Los
Angeles, CA, USA
5. Department of Family Medicine and Public Health Sciences, School of Medicine, Karmanos
Cancer Institute, Wayne State University, Detroit, MI, USA
6. Joseph J. Zilber School of Public Health, University of Wisconsin-Milwaukee, Milwaukee,
WI, USA
7. Departments of Medicine and Pathology, Medical College of Wisconsin, Milwaukee, WI,
USA

77
4.1 Abstract
Background: Evidence of an association between tobacco exposure and breast cancer (BC) risk
has been inconsistent. Studies among premenopausal women suggest smoking is associated with
an increased risk for BC, but few studies have evaluated associations between lifetime smoking
and BC subtype risk, or whether associations differ by race and socioeconomic position (SEP).
Methods: We evaluated the association between lifetime personal smoking and BC risk, and
potential interactions by race and SEP, using data from the Young Women’s Health History
(YWHHS) study. YWHHS is a population-based case-control study of non-Hispanic Black and
White women 20-49 years of age (1,812 cases, 1,381 controls). Smoking and BC risk was
examined by BC subtype using sample-weighted polytomous logistic regression.
Results: We observed heterogeneity in risk associated with ever smoking at least one
cigarette/day for at least six months compared to never smoking by BC subtype
(Pheterogeneity=0.02). Ever smoking was significantly associated with an increased risk of Luminal
A BC (adjusted odds ratio (aOR) 1.33; 95% confidence interval (CI) 1.06-1.68) and HER2-type
BC (aOR 1.96; 95% CI 1.22-3.15) compared to never smoking but was not associated with
Luminal B or Triple Negative subtypes. No significant interactions were observed for overall BC
by race. Risk of BC differed for current and former smokers by SEP (Pinteraction=0.02).
Conclusion: Smoking was associated with an increased risk for BC in young women.
Impact: Findings confirm prior reports of a positive association between smoking and Luminal
A BC and identify a novel association between smoking and HER2-type BC risk.

4.2 Introduction

78
Breast cancer (BC) is the most commonly diagnosed cancer among young women (under
age 50 years) with an annual incidence in the United States (US) of 126.4 per 100,000 persons
and an average annual percent increase of 0.4% per year in 2011-2015 (1; 2). Breast cancer is
also recognized to have different etiologies by molecular subtypes which have been categorized
by estrogen receptor (ER), progesterone receptor (PR), Human Epidermal Growth Factor
Receptor 2 (HER2) status and tumor grade as: Luminal A, Luminal B, HER2-type, and Triple
Negative BC (TNBC) (3). Evidence suggests lifetime smoking is associated with an increased
risk of BC, but the association between lifetime smoking and BC risk for different tumor
subtypes has not been well studied among young women (4-8).
Racial and socioeconomic inequities in BC incidence also persist in the United States (2;
9-14). Among young women, non-Hispanic White (NHW) women have the highest incidence of
Luminal A BC, the subtype associated with the highest survival, with an annual incidence of
44.3 per 100,000 (compared to 35.8 among non-Hispanic Black (NHB) women) in 2011-2013
(15). Conversely, young NHB women have the highest incidence of TNBC, which is associated
with the poorest prognosis (17.5 compared to 9.3 per 100,000 among NHW women in 2011-
2013) (15). Research has shown that the incidence of hormone receptor (HR)-positive BC
increases with increasing socioeconomic position (SEP), but that there is potentially an inverse
association with hormone receptor (HR)-negative BC (16; 17). Socioeconomic position is an
indicator of the social and economic factors relating to an individual’s resources and exposures
that can affect their health (18). Little research has evaluated whether associations between
lifetime smoking and BC risk overall and by tumor subtypes varies by race or SEP.
Recent studies have suggested that active smoking is associated with BC risk in young
women (4-8). In the Nurses’ Health Study (NHS), lifetime smoking among premenopausal

79
women was associated with an 11% increase in BC risk for each additional 20 pack-years of use
(hazard ratio 1.11; 95% CI 1.07-1.15) (6). In the Black Women’s Health Study (BWHS), risk of
premenopausal BC was 70% higher (incidence rate ratio 1.70; 95% CI 1.05–2.75) among those
who initiated smoking before age 18 and smoked at least 20 pack-years compared to never
smokers (7). Furthermore, Morabia et.al. reported that among premenopausal women, ever
smokers were more than twice as likely to have ER+ BC irrespective of the amount they smoked
per day (19).
Differences in smoking behaviors have also been described by race and SEP (20-22). The
Centers for Disease Control and Prevention (CDC) reports that Black adults tend to initiate
smoking at a later age but have lower smoking cessation rates compared to White smokers (20;
21). Population-based surveys have consistently reported a higher prevalence of smoking and
heavier smoking among NHW women compared to NHB women (22). Also, poorer women tend
to smoke more cigarettes per day than wealthier women (22).
In this study we build on previous research by evaluating the association between
cumulative lifetime personal smoking and BC risk, overall and by tumor subtype, in a
socioeconomically diverse population-based study of young NHB and NHW women. We further
evaluate whether smoking-related risk is modified by race or SEP.

4.3 Methods
4.3.1. Source population of cases and controls
Data are from a population-based case-control study of BC among young women, the
Young Women’s Health History Study (YWHHS). Briefly, case participants were identified
from residents of the Los Angeles County (LA) and Metropolitan Detroit Surveillance,

80
Epidemiology, and End Results (SEER) Registry areas. Women diagnosed with histologically
confirmed incident invasive primary BC between 2010 to 2015 were eligible. Women diagnosed
with breast lymphoma, Paget’s disease, mesenchymal tumors, including sarcomas and
hemangiosarcomas of the breast were excluded from the study. A total of 5,309 patients meeting
the case definition were identified by the two above-mentioned SEER registries (2,782 LA,
2,527 Detroit). Eighty percent of identified cases were sampled for recruitment (n=4,269) and of
these, 236 were deemed ineligible before screening, 707 were deemed ineligible after screening,
82 were not contacted because their physician or hospital did not provide permission for contact,
147 had died or were too ill to be interviewed, 70 no longer lived in the study area, 177 were not
located, 623 patients declined to be interviewed, and 415 were lost to follow-up or contact could
not be established. In total, 1,812 women with invasive BC (1,130 NHW, 682 NHB) completed
an in-person interview (response rate 59.8%).
Area-based controls were sampled from over 24,000 postal addresses based on the 2010
US Census and frequency matched to cases on race, study region, and five-year age group.
Among selected postal addresses, potentially eligible household members were identified,
screened for eligibility, and asked in-person to participate in the study. Overall, 3,414 potential
control participants were identified, 2720 completed screening interviews and 1,988 were
deemed potentially eligible. Of these, 55 women declined further contact during the screening
interview, 12 were deemed ineligible after screening, 4 had died or were too ill to be
interviewed, 30 no longer lived in the study area, 6 were not located, 327 patients declined to be
interviewed, and 173 were lost to follow-up or contact could not be established. In total, 1,381
control participants (716 NHW, 665 NHB) completed the study interview (response rate 53%).

81
All participants met the following criteria: female, self-identified race/ethnicity as NHB
or NHW, aged 20-49 years at reference date, resident of Los Angeles County or the tri county
(Macomb, Oakland and Wayne) metropolitan Detroit area, born in the US, no previous BC
diagnosis (invasive or in situ), no previous cancer diagnosis except for cervical in situ or non-
melanoma skin cancer before the reference date, able to complete the interview in English, not
institutionalized, and physically and mentally able to complete the study interview. The study
reference date was defined as the date of histologically confirmed BC diagnosis for cases and the
date of the screening interview minus four months for controls.
Institutional review boards at the University of Southern California, the California
Committee for the Protection of Human Subjects, the University of Wisconsin – Milwaukee,
Michigan State University, Wayne State University, the Michigan Department of Community
Health, the Medical College of Wisconsin and the California Cancer Registry approved the
study. Written informed consent was completed by each participant before study interviews were
conducted.

4.3.2 Tumor Subtyping
Information on tumor characteristics was collected routinely in both SEER tumor
registries and was available for BC cases included in this study. Specifically, hormone receptor
(ER/PR) status, HER2 status, and tumor grade for subtyping were included for tumor subtyping.
Molecular subtypes were categorized as Luminal A (ER/PR+, HER2-, grade 1/2), Luminal B
(ER/PR+, HER2+ or HER2-, grade 3+), HER2-type (ER-, PR-, HER2+), and TNBC (ER-, PR-
HER2-) (3).

82
4.3.3 Data collection - lifestyle factors and tobacco exposures
Structured in-person interviews were conducted with participants in the two study areas
by interviewers specifically trained for this study. Sociodemographic characteristics included
age, place of birth, duration of residence in LA/Detroit, race/ethnicity, education, and household
income in the 12 months before reference date and the number of household members supported
by that income. Participants were asked about established and suspected breast cancer risk
factors including: age at menarche; number of pregnancies and outcome of each pregnancy;
height and weight in childhood, adolescence and adulthood; family history of breast cancer; and
menopausal status. Personal lifetime smoking histories were collected using structured questions
about smoking status, and open-ended questions about age at initiation, periods of smoking
cessation, and the average number of cigarettes smoked per day.

4.3.4 Statistical analysis
Sociodemographic variables include site (Metropolitan Detroit or LA County), race (self-
reported NHB or NHW), age at diagnosis/reference year, education, and the household poverty
level (HHP) as a measure of SEP. Race was defined as Black/African American, White, Asian,
Native Hawaiian/other Pacific Islander, American Indian/Alaska Native, or Other. Ethnicity was
categorized as Hispanic of Latina origin or not. HHP was derived from self-reported gross
income 12 months before reference date and the number of members supported by that income in
the household to calculate the household poverty percent of federal poverty level (FPL) (≥200%
of FPL, <200% of FPL). The values were calculated using the 2009-2014 poverty thresholds and
the cut point was chosen to define a dichotomous variable that would be based on a broader
definition of low SEP compared to what is typically used to qualify for federal aid (23; 24).

83
Established BC risk factors were assessed then categorized as follows: first degree family
history of BC (yes, no, missing/don’t know), age at menarche (≤11, 12, 13, ≥14 years), lifetime
cumulative alcohol use (0, 0.1-6.9, 7-13.9, 14-27.9 and ≥28 grams/day), body mass index (BMI)
12 months before reference date (underweight: <18.5, normal: 18.5-24.9, overweight:25.0-29.9,
obese: ≥30 kg/m
2
), menopausal status (premenopausal, peri- & postmenopausal), and a joint
variable of parity and age at first full-term pregnancy (FFTP) (nulliparous, 1-2 children and <25
years, 1-2 children and ≥25, 3+ children and <25 years, 3+ children and ≥25).
Ever smoking was defined as women who smoked at least one cigarette a day for six
months or more. Personal smoking status was described as never, former, and current smokers.
Former smokers were categorized as quitting ≥10 or <10 years before the study reference date,
based on the distribution of the data. Lifetime smoking intensity in pack-years was calculated
from the number of cigarettes smoked per day divided by 20 cigarettes per pack and multiplied
by years of smoking history (never smoker, <5, 5-19, ≥20 pack-years). We also evaluated the
timing of smoking initiation in relation to FFTP among parous women (never smoker, after
FFTP, before FFTP), age at smoking initiation (never smoker, <18, 18-24, ≥25 years), and time
since smoking initiation (never smoker, <20, 20-29, ≥30 years). Last, we assessed the average
number of cigarettes smoked per day (never smoker, <5, 5-19, ≥20 cigarettes per day).
Comparisons of matching characteristics and lifestyle factors were made by personal
smoking status using a sample-weighted chi-square test for categorical variables and sample-
weighted Wald’s test for continuous variables. BC relative risk was estimated by odds ratio and
assessed, overall and by BC subtype, in crude and multivariable adjusted models. Sample-
weighted multivariable logistic regression was used to estimate the adjusted odds ratios (aOR)
and 95% confidence intervals (CIs) for the association between lifetime cigarette smoking

84
characteristics and BC risk overall and polytomous logistic regression for associations by BC
subtype. Sample weights take into account the sampling design and non-response. For case-
control analyses, sample weights take into account the case population; for control-only
demographic summaries, controls were weighted to each study area population based on the
2010 US Census. Control status was the reference outcome for all analyses.
Models were adjusted for site, age, HHP, first degree family member with BC, BMI 12
months before reference date, alcohol use, joint parity/age at FFTP, and menopausal status based
on assessments for confounding with ever smoking status and overall BC status or based on a
priori knowledge. Wald’s tests were employed to assess heterogeneity in the odds ratio estimates
by BC subtypes. Analyses were also conducted stratified by race and HHP; cross-product
interaction terms of smoking exposures by each stratum were evaluated by Wald’s test.
Ever/never smoking status was missing for 136 participants (87 cases and 49 controls)
who were excluded from the analysis. Missing data was captured as a separate category for all
other smoking characteristics but did not contribute to the estimated p-values. Breast cancer
subtype was missing for 125 cases who were not included in analyses by breast cancer subtype.
For all covariates included in multivariable analyses, values were imputed for missing data
except first degree family member with BC, where missing or “don’t know” (n=38) were
combined as a separate category. Missing values for the covariates alcohol use (n=20), BMI 12
months before reference date (n=9), and the joint indicator for parity/age at FFTP (n=1), were
imputed based on the sample median for race, HHP, and joint parity/age at FFTP. If participants
were also missing information for HHP or joint parity/age at FFTP, the overall study median was
used to impute missing values. Missingness was not associated with race, HHP or case status, but

85
there was a higher proportion of participants with responses missing for a first-degree family
member with BC if HHP was also missing (P=0.02).
Tests were assessed at a significance level of P<0.05, with interaction terms assessed at a
level of Pinteraction<0.10. Analyses were conducted using Stata version 14.2 (StataCorp LLC,
College Station, TX).

4.4 Results
The demographic characteristics of ever, former, and never smokers among controls
(n=1,381) are presented in Table 4.1. We observed a greater proportion of former and current
smokers in Metropolitan Detroit compared to LA County (P=0.001). On average, former and
current smokers were older than never smokers, and there was a higher proportion of former
smokers among NHW women and a higher proportion of current smokers among women with an
HHP <200% of the FPL (P<0.001 for each factor). A higher proportion of former and current
smokers were peri- and postmenopausal (P=0.002). Smoking status differed by the joint
parity/age at FFTP status with high proportions of FFTP after age 25 years among former
smokers and high proportions of FFTP before age 25 year for current smokers (P<0.001). The
proportion of former and current smoking increased with higher alcohol use (P<0.001). Smoking
status did not differ by BMI, age at menarche or by history of BC among a first degree relative.
Table 4.2 shows a summary of smoking characteristics for controls (n=1, 332) and all
cases (n=1,725) and by BC subtype. Cases compared to controls were more likely to be former
smokers (21.2% vs. 17.9%), to have initiated smoking before FFTP (33.7% vs. 30.2%), and to
have initiated smoking at ages under 18 years (23.5% vs. 21.1%) or at ages ≥25 years (4.3 vs.

86
2.8). The highest proportions with greater tobacco exposure were consistently observed among
cases with HER2-type BC.
In multivariable adjusted models, a positive nonsignificant association was identified
between ever smoking and overall BC risk (aOR 1.20; 95% CI 0.99-1.46) (Table 4.3). However,
significant heterogeneity was observed in the association between ever smoking and BC risk by
BC subtype (Pheterogeneity=0.02). We identified a significantly increased risk of Luminal A BC
(aOR 1.33; 95% CI 1.06-1.68) and HER2-type BC (aOR 1.96; 95% CI 1.22-3.15) for ever
smokers compared to never smokers, whereas ever smoking was not significantly associated
with the Luminal B or TNBC subtypes. These increased risks of Luminal A and HER2-type BC
were also seen among former smokers compared to never smokers, while current smoking was
not statistically significantly associated with Luminal A or HER2-type BC. Increasing lifetime
smoking pack-years was associated with a significantly increased risk of Luminal A BC and
HER2-type BC (Ptrend=0.02 for both).
We also observed a significant increase in overall BC risk among women who initiated
smoking before their FFTP (aOR 1.25; 95% CI 1.02-1.88) (Table 4.3). Smoking before FFTP
was associated with an increased risk of Luminal A BC (aOR 1.43; 95% CI 1.09-1.88) and,
though not statistically significant, an increase in risk for HER2-type (aOR 1.74; 95% CI 0.97-
3.14). Increasing age at smoking initiation was significantly associated with an increasing risk of
HER2-type BC (Ptrend=0.04). Age at initiation was not statistically significantly associated with
Luminal A BC by trend analysis (Ptrend=0.07), however, initiating at ages ≥25 years doubled the
risk of Luminal A BC (aOR 2.22; 95% CI 1.29-3.85). Longer time since initiation and a higher
number of cigarettes per day were both associated with an increased risk of Luminal A and
HER2-type BC (Ptrend<0.05 for each).

87
We did not find a significant difference in the observed association between each of the
smoking characteristics of interest and case status, by race (Pinteraction>0.05 for all) (Table 4.4).
However, among NHW women, with reference category of never smokers, a significantly
increased risk was observed for ever smoking (aOR 1.37; 95% CI 1.04-1.82), current smoking
(aOR 1.43; 95% CI 1.00-2.04) and smoking initiated before FFTP (aOR 1.44; 95% CI 1.06-
1.96). Also, an increased risk (aOR 1.58; 95% CI 1.16-2.16) was observed for initiating smoking
before age 18. None of these significant associations for NHW women were observed among
NHB women.
We observed a significant difference in personal smoking status and overall BC risk by
HHP (Pinteraction=0.03) (Table 4.4). Among women with an HHP level ≥200%, current smokers
had a higher BC risk (aOR 1.54; 95% CI 1.02-2.32) compared to never smokers, but risk was not
increased among former smokers. In contrast, among women with an HHP level <200%, former
smokers had an increased risk of BC (aOR 1.73; 95% CI 1.10-2.73), while current smoking was
not significantly associated with risk compared to never smokers. Initiating smoking before age
18 or after age 25 was associated with an increased risk of BC among women with an HHP level
≥200% (aOR 1.46; 95% CI 1.02-2.08 and aOR 2.65; 95% CI 1.19-5.92, respectively) although
this did not significantly differ from women with an HHP level <200% (Pinteraction=0.42).

4.5 Discussion
We identified an association between several lifetime smoking characteristics and BC
risk, overall and by BC subtype, in the Young Women’s Health History Study. Specifically, we
observed a 20% increased risk of BC with lifetime smoking (ever vs. never smoker) and a 33%
increased risk for Luminal A and 96% increased risk for HER2-type BC. A higher magnitude of

88
risk was suggested among NHW women compared to NHB women for most of the smoking
characteristics. Also, overall BC risk for current and former smokers differed by HHP where
current smokers had the highest risk of breast cancer among women with an HHP level ≥200%
(aOR=1.54) and former smokers had the highest risk among poorer women (aOR=1.73 for HHP
<200% of FPL).
Previous studies have evaluated the association between smoking and BC risk by
subtype, and many have identified a positive association between smoking and risk of Luminal A
or hormone receptor positive BC, however, only two previous studies have been conducted
among younger or premenopausal women (6; 19; 25-29). Risk for BC has been shown to vary by
age or menopausal status for several factors, and given the different hormonal milieu, may also
vary for smoking status (30-32). Our findings are consistent with the increased risk of BC found
among premenopausal smokers in both previous studies, but especially with the findings of
Kawai et.al. (2014) who identified a 40% increased risk of ER+ BC associated with ever
smoking among a largely white population of women aged 20-44 years (19; 28). Also, our
finding for Luminal A BC (which is ER+) was consistent with their report of a 40% increased
risk of ER+ BC when smoking was initiated before FFTP. However, we identified an increased
risk of BC with increasing pack-years of smoking, in contrast to their null finding (Ptrend=0.90)
which may be related, in part, to different cut points for smoking pack-years (never, <2.5, 2.5-
4.9, 5-9.9, 10-14.9, ≥15 pack-years). We observed higher proportions with moderate smoking (5-
19 pack-years) and were sufficiently powered to detect a significant trend in our population
(Ptrend=0.02).
In addition to Luminal A BC, we found that ever smoking was associated with a 96%
higher risk of HER2-type BC. Three previous BC studies on smoking have included an

89
assessment of risk pattern in the HER2-subtype but none, to our knowledge, have found an
association between smoking and HER2-subtype in any age group (33-35). Inclusion of HER2-
type BC in population-based epidemiologic studies is still a relatively new and evolving field as
HER2 protein expression was often underreported in the pathology reports of cases diagnosed
before 2005 and routine reporting of HER2 status was not available in SEER cancer registries
until 2010 (2; 36).
Our finding of a strong association between former smoking and BC risk was
unexpected, but there may be some biological explanations for this observation. Although
smoking is known to be associated with known BC risk factors, including high alcohol
consumption and early menarche which, in turn, are associated with higher endogenous estrogen
levels; current smoking premenopausal women in the NHS displayed altered estrogen
metabolism and had significantly lower estradiol levels compared to never smokers (32; 37; 38).
Therefore, the lack of an association among current smokers may be due to some protective
effects conferred by estrogen suppression among smokers. Still, the elevated risk among former
smokers may suggest a latent toxicity of the carcinogens in tobacco on breast tissue (5; 39; 40).
On the other hand, we detected a significant difference in personal smoking status by
HHP where current smoking was associated with an elevated BC risk among wealthier women
with an HHP level ≥200%. Women of high SEP may experience higher BC risk because of
lower parity and older age at FFTP; though we adjusted for these potential associations in our
analyses, there may be residual confounding (16). In this population we observed a higher
prevalence of nulliparous women and women with FFTP at ages above 25 years among women
with HHP ≥200%. The observed differences in the distribution of nulliparous women and those
with FFTP above 25 years of age, by HHP, may suggest that without the protective effects of

90
reproductive factors, such as younger age at FFTP, current smoking is associated with increased
BC risk among young women. This association warrants further confirmation.
Although we did not detect a significant statistical interaction for any of the smoking
characteristics with race, we observed several characteristics that were significantly associated
with smoking among NHW participants. Smoking was more common among NHW women with
BC compared to NHB women (Table 4.1), so the significant associations identified may be
driven by the higher prevalence of smoking among the NHW women in our population.
This study had many strengths. A large sample of NHB and NHW younger cases were
identified in two socioeconomically diverse population-based SEER registry areas. Bias in
assessing cancer diagnoses was minimized by using SEER registry data for case ascertainment
and BC subtyping. An extensive area population-based sampling approach was utilized to
identify and recruit controls, and sample weights were estimated and applied to the statistical
models to account for sampling and nonresponse bias. This study represents one of only three
studies to explore the association between smoking and BC subtype with a sample of HER2-type
BCs among a population of young women.
Limitations include concerns about differential misclassification of exposures that may
occur in a case-control study based on self-reported recall, however lifetime history calendars
were used to prompt memory and applied equally to cases and controls. We anticipate that any
misclassification of smoking exposures is likely nondifferential and would thus lead to an
underestimation of the true association in the population. The response rates observed may be
deemed as a limitation for this study, but they are similar to rates observed for other studies and
especially given the data collection period, target age group, and anticipated challenges of
recruiting disadvantaged populations (41-45). Furthermore, enumeration of all cases and 86% of

91
sampled controls in the Los Angeles County and Metropolitan Detroit SEER registry areas
allowed for the incorporation of non-response sample weights, mitigating this limitation.
Another limitation is a lack of information on other combustible tobacco products in addition to
cigarette use. Thus, there was potential for differential misclassification of intensity of tobacco
exposure if BC cases are more likely to be polytobacco users. In a recent review of trends in use
of little cigars and cigarillos, Messer et.al. found that about 70% of female adolescent and young
adult cigar smokers also smoked cigarettes in 2002-2011 (46). Therefore, our analysis on
cigarette use likely captured a substantial degree of the tobacco exposure in this population with
some potential for underestimating the true association. Unmeasured confounding may
contribute to some bias in the results, as in all observational studies.
Our results indicate that smoking, including ever smoking as well as heavier and longer-
term smoking before FFTP, was associated with an increased risk for BC overall and for
Luminal A and HER2-type BC in young women. A higher risk of BC was suggested among
NHW women and we detected significant differences in BC risk for former and current smokers
by HHP where risk was highest for poorer women that were former smokers and wealthier
women that were current smokers. Further research to uncover the biological mechanisms for an
increased risk of the Luminal A and HER2-type BC due to cigarette smoking is warranted. Also,
as new combustible tobacco products are being developed and used in populations of young
people, future studies should investigate the effects of these newer products on BC risk. In sum,
based on these findings and consistent with other studies, education in smoking cessation and
efforts in preventing smoking initiation early in life are necessary to reduce BC incidence among
young women in the US.

92
4.6 References
1. Ward, E. M., Sherman, R. L., Henley, S. J., Jemal, A., Siegel, D. A., Feuer, E. J., . . .
Cronin, K. A. (2019). Annual Report to the Nation on the Status of Cancer, Featuring
Cancer in Men and Women Age 20-49 Years. J Natl Cancer Inst, 111(12), 1279-1297.
doi:10.1093/jnci/djz106
2. DeSantis, C. E., Ma, J., Gaudet, M. M., Newman, L. A., Miller, K. D., Goding Sauer, A., . .
. Siegel, R. L. (2019). Breast cancer statistics, 2019. CA Cancer J Clin, 69(6), 438-451.
doi:10.3322/caac.21583
3. Provenzano, E., Ulaner, G. A., & Chin, S. F. (2018). Molecular Classification of Breast
Cancer. PET Clin, 13(3), 325-338. doi:10.1016/j.cpet.2018.02.004
4. Reynolds, P. (2013). Smoking and breast cancer. J Mammary Gland Biol Neoplasia, 18(1),
15-23. doi:10.1007/s10911-012-9269-x
5. Kispert, S., & McHowat, J. (2017). Recent insights into cigarette smoking as a lifestyle risk
factor for breast cancer. Breast Cancer (Dove Med Press), 9, 127-132.
doi:10.2147/BCTT.S129746
6. Xue, F., Willett, W. C., Rosner, B. A., Hankinson, S. E., & Michels, K. B. (2011).
Cigarette smoking and the incidence of breast cancer. Arch Intern Med, 171(2), 125-133.
doi:10.1001/archinternmed.2010.503
7. Rosenberg, L., Boggs, D. A., Bethea, T. N., Wise, L. A., Adams-Campbell, L. L., &
Palmer, J. R. (2013). A prospective study of smoking and breast cancer risk among
African-American women. Cancer Causes Control, 24(12), 2207-2215.
doi:10.1007/s10552-013-0298-6
8. Morabia, A., Bernstein, M. S., Bouchardy, I., Kurtz, J., & Morris, M. A. (2000). Breast
cancer and active and passive smoking: the role of the N-acetyltransferase 2 genotype. Am
J Epidemiol, 152(3), 226-232.
9. Danforth, D. N., Jr. (2013). Disparities in breast cancer outcomes between Caucasian and
African American women: a model for describing the relationship of biological and
nonbiological factors. Breast Cancer Res, 15(3), 208. doi:10.1186/bcr3429
10. National Cancer Institute Surveillance Research Program. National Cancer Institute
SEER*Stat software. Retrieved September 15, 2019 https://seer.cancer.gov/seerstat
version 8.3.6
11. Liu, L., Deapen, D., & Bernstein, L. (1998). Socioeconomic status and cancers of the
female breast and reproductive organs: a comparison across racial/ethnic populations in
Los Angeles County, California (United States). Cancer Causes Control, 9(4), 369-380.
12. Borugian, M. J., Spinelli, J. J., Abanto, Z., Xu, C. L., & Wilkins, R. (2011). Breast cancer
incidence and neighbourhood income. Health Rep, 22(2), 7-13.

93
13. Vainshtein, J. (2008). Disparities in breast cancer incidence across racial/ethnic strata and
socioeconomic status: a systematic review. J Natl Med Assoc, 100(7), 833-839.
14. Vona-Davis, L., & Rose, D. P. (2009). The influence of socioeconomic disparities on
breast cancer tumor biology and prognosis: a review. J Womens Health (Larchmt), 18(6),
883-893. doi:10.1089/jwh.2008.1127
15. Shoemaker, M. L., White, M. C., Wu, M., Weir, H. K., & Romieu, I. (2018). Differences in
breast cancer incidence among young women aged 20-49 years by stage and tumor
characteristics, age, race, and ethnicity, 2004-2013. Breast Cancer Res Treat.
doi:10.1007/s10549-018-4699-9
16. Akinyemiju, T. F., Pisu, M., Waterbor, J. W., & Altekruse, S. F. (2015). Socioeconomic
status and incidence of breast cancer by hormone receptor subtype. Springerplus, 4, 508.
doi:10.1186/s40064-015-1282-2
17. Andaya, A. A., Enewold, L., Horner, M. J., Jatoi, I., Shriver, C. D., & Zhu, K. (2012).
Socioeconomic disparities and breast cancer hormone receptor status. Cancer Causes
Control, 23(6), 951-958. doi:10.1007/s10552-012-9966-1
18. Galobardes, B., Shaw, M., Lawlor, D. A., Lynch, J. W., & Davey Smith, G. (2006).
Indicators of socioeconomic position (part 1). J Epidemiol Community Health, 60(1), 7-12.
doi:10.1136/jech.2004.023531
19. Morabia, A., Bernstein, M., Ruiz, J., Heritier, S., Diebold Berger, S., & Borisch, B. (1998).
Relation of smoking to breast cancer by estrogen receptor status. Int J Cancer, 75(3), 339-
342.
20. Roberts, M. E., Colby, S. M., Lu, B., & Ferketich, A. K. (2016). Understanding Tobacco
Use Onset Among African Americans. Nicotine Tob Res, 18 Suppl 1, S49-56.
doi:10.1093/ntr/ntv250
21. Centers for Disease Control and Prevention Office on Smoking and Health. (2019).
Smoking & Tobacco Use: African Americans and Tobacco Use.
https://www.cdc.gov/tobacco/disparities/african-americans/index.htm
22. Shavers, V. L., Fagan, P., Alexander, L. A., Clayton, R., Doucet, J., & Baezconde-
Garbanati, L. (2006). Workplace and home smoking restrictions and racial/ethnic variation
in the prevalence and intensity of current cigarette smoking among women by poverty
status, TUS-CPS 1998-1999 and 2001-2002. J Epidemiol Community Health, 60 Suppl 2,
34-43. doi:10.1136/jech.2006.046979
23. United States Census Bureau. (2020, August 21, 2020). Poverty Thresholds. Retrieved
from https://www.census.gov/data/tables/time-series/demo/income-poverty/historical-
poverty-thresholds.html
24. Fass, S. (2009, May 2009). Measuring Poverty in the United States. Retrieved from
http://www.nccp.org/publication/measuring-poverty-in-the-united-states/

94
25. Butler, E. N., Tse, C. K., Bell, M. E., Conway, K., Olshan, A. F., & Troester, M. A. (2016).
Active smoking and risk of Luminal and Basal-like breast cancer subtypes in the Carolina
Breast Cancer Study. Cancer Causes Control, 27(6), 775-786. doi:10.1007/s10552-016-
0754-1
26. Ellingjord-Dale, M., Vos, L., Vik Hjerkind, K., Hjartaker, A., Russnes, H. G., Tretli, S., . . .
Ursin, G. (2018). Number of Risky Lifestyle Behaviors and Breast Cancer Risk. JNCI
Cancer Spectr, 2(3), pky030. doi:10.1093/jncics/pky030
27. Kabat, G. C., Kim, M., Phipps, A. I., Li, C. I., Messina, C. R., Wactawski-Wende, J., . . .
Rohan, T. E. (2011). Smoking and alcohol consumption in relation to risk of triple-
negative breast cancer in a cohort of postmenopausal women. Cancer Causes Control,
22(5), 775-783. doi:10.1007/s10552-011-9750-7
28. Kawai, M., Malone, K. E., Tang, M. T., & Li, C. I. (2014). Active smoking and the risk of
estrogen receptor-positive and triple-negative breast cancer among women ages 20 to 44
years. Cancer, 120(7), 1026-1034. doi:10.1002/cncr.28402
29. Nishino, Y., Minami, Y., Kawai, M., Fukamachi, K., Sato, I., Ohuchi, N., & Kakugawa, Y.
(2014). Cigarette smoking and breast cancer risk in relation to joint estrogen and
progesterone receptor status: a case-control study in Japan. Springerplus, 3, 65.
doi:10.1186/2193-1801-3-65
30. Endogenous Hormones Breast Cancer Collaborative Group, Key, T. J., Appleby, P. N.,
Reeves, G. K., Roddam, A. W., Helzlsouer, K. J., . . . Strickler, H. D. (2011). Circulating
sex hormones and breast cancer risk factors in postmenopausal women: reanalysis of 13
studies. Br J Cancer, 105(5), 709-722. doi:10.1038/bjc.2011.254
31. Endogenous Hormones Breast Cancer Collaborative Group, Key, T. J., Appleby, P. N.,
Reeves, G. K., Travis, R. C., Alberg, A. J., . . . Vineis, P. (2013). Sex hormones and risk of
breast cancer in premenopausal women: a collaborative reanalysis of individual participant
data from seven prospective studies. Lancet Oncol, 14(10), 1009-1019. doi:10.1016/S1470-
2045(13)70301-2
32. Verkasalo, P. K., Thomas, H. V., Appleby, P. N., Davey, G. K., & Key, T. J. (2001).
Circulating levels of sex hormones and their relation to risk factors for breast cancer: a
cross-sectional study in 1092 pre- and postmenopausal women (United Kingdom). Cancer
Causes Control, 12(1), 47-59.
33. Ellingjord-Dale, M., Vos, L., Hjerkind, K. V., Hjartaker, A., Russnes, H. G., Tretli, S., . . .
Ursin, G. (2017). Alcohol, Physical Activity, Smoking, and Breast Cancer Subtypes in a
Large, Nested Case-Control Study from the Norwegian Breast Cancer Screening Program.
Cancer Epidemiol Biomarkers Prev, 26(12), 1736-1744. doi:10.1158/1055-9965.EPI-17-
0611
34. Baglia, M. L., Cook, L. S., Mei-Tzu, C., Wiggins, C., Hill, D., Porter, P., & Li, C. I.
(2018). Alcohol, smoking, and risk of Her2-overexpressing and triple-negative breast

95
cancer relative to estrogen receptor-positive breast cancer. Int J Cancer, 143(8), 1849-
1857. doi:10.1002/ijc.31575
35. Turkoz, F. P., Solak, M., Petekkaya, I., Keskin, O., Kertmen, N., Sarici, F., . . . Altundag,
K. (2013). Association between common risk factors and molecular subtypes in breast
cancer patients. Breast, 22(3), 344-350. doi:10.1016/j.breast.2012.08.005
36. Surveillance, E., and End Results (SEER) Program,. Breast Subtype (2010+). Retrieved
from https://seer.cancer.gov/seerstat/databases/ssf/breast-subtype.html
37. Gu, F., Caporaso, N. E., Schairer, C., Fortner, R. T., Xu, X., Hankinson, S. E., . . . Ziegler,
R. G. (2013). Urinary concentrations of estrogens and estrogen metabolites and smoking in
caucasian women. Cancer Epidemiol Biomarkers Prev, 22(1), 58-68. doi:10.1158/1055-
9965.EPI-12-0909
38. Reynolds, P., Hurley, S., Goldberg, D. E., Anton-Culver, H., Bernstein, L., Deapen, D., . . .
Ziogas, A. (2004). Active smoking, household passive smoking, and breast cancer:
evidence from the California Teachers Study. J Natl Cancer Inst, 96(1), 29-37.
39. Petrakis, N. L., Gruenke, L. D., Beelen, T. C., Castagnoli, N., Jr., & Craig, J. C. (1978).
Nicotine in breast fluid of nonlactating women. Science, 199(4326), 303-305.
40. Verde, Z., Santiago, C., Chicharro, L. M., Reinoso-Barbero, L., Tejerina, A., Bandres, F.,
& Gomez-Gallego, F. (2016). Effect of Genetic Polymorphisms and Long-Term Tobacco
Exposure on the Risk of Breast Cancer. Int J Mol Sci, 17(10). doi:10.3390/ijms17101726
41. Marchbanks, P. A., McDonald, J. A., Wilson, H. G., Burnett, N. M., Daling, J. R.,
Bernstein, L., . . . Spirtas, R. (2002). The NICHD Women's Contraceptive and
Reproductive Experiences Study: methods and operational results. Ann Epidemiol, 12(4),
213-221.
42. Xu, M., Richardson, L., Campbell, S., Pintos, J., & Siemiatycki, J. (2018). Response rates
in case-control studies of cancer by era of fieldwork and by characteristics of study design.
Ann Epidemiol, 28(6), 385-391. doi:10.1016/j.annepidem.2018.04.001
43. Palmer, J. R., Ambrosone, C. B., & Olshan, A. F. (2014). A collaborative study of the
etiology of breast cancer subtypes in African American women: the AMBER consortium.
Cancer Causes Control, 25(3), 309-319. doi:10.1007/s10552-013-0332-8
44. Pinn;, V., Harden;, J., Brawley;, O., Blehar;, M., McGowan;, J., Melandez-Bohler;, G., &
Washington, C. (2003). Outreach Notebook: For the inclusion, recruitment and retention
of women and minority subjects in clinical research. National Institutes of Health.
45. Bartlett, D. W. R. (2013). Recruitment and Retention of African American and Hispanic
Girls and Women in Research. Public Health Nurs, 30(2), 159-166. doi:10.1111/phn.12014

96
46. Messer, K., White, M. M., Strong, D. R., Wang, B., Shi, Y., Conway, K. P., & Pierce, J. P.
(2015). Trends in use of little cigars or cigarillos and cigarettes among U.S. smokers, 2002-
2011. Nicotine Tob Res, 17(5), 515-523. doi:10.1093/ntr/ntu179

97
Table 4.1. Characteristics of the control participants in the Young Women’s Health History
Study by personal smoking status (n=1,381)
a

Total Population Never Smoker Former Smoker Current Smoker P-value
N (W%)
a
1332 (100%)
b
853 (66.5%) 207 (14.3%) 272 (19.3%)

Study Site

<0.001
Metropolitan Detroit 700 (60.7) 417 (61.2) 109 (14.9) 174 (23.9)

Los Angeles County 632 (39.4) 436 (74.7) 98 (13.3) 98 (12.1)

Age at reference year, years
(weighted mean (weighted 95% CI))
34.3 (33.2-35.4) 33.4 (32.0-34.7) 37.1 (35.4-38.9) 35.4 (33.9-36.8) 0.001
Age at reference year, years

0.003
20-29 238 (36.6) 176 (75.3) 16 (7.1) 46 (17.6)

30-39 466 (30.1) 294 (62.1) 84 (18.8) 88 (19.0)

40-49 628 (33.3) 383 (60.7) 107 (18.0) 138 (21.2)

Race

<0.001
Non-Hispanic White 679 (64.5) 401 (63.3) 156 (18.7) 122 (18.0)

Non-Hispanic Black 653 (35.5) 452 (72.2) 51 (6.3) 150 (21.5)

Household poverty level

<0.001
≥200% of federal poverty level 691 (53.4) 479 (72.2) 137 (17.2) 75 (10.6)

<200% of federal poverty level 601 (43.7) 352 (59.3) 66 (11.2) 183 (29.5)

Missing 40 (2.9) 22 (68.9) 4 (7.3) 14 (23.7)

Education

<0.001
High school diploma or less 286 (18.3) 162 (57.0) 30 (11.4) 94 (31.7)

Vocational school, associate's
degree, or some college
541 (43.4) 317 (61.6) 92 (15.6) 132 (22.8)

Bachelor's degree or higher 505 (38.3) 374 (76.6) 85 (14.1) 46 (9.3)

Body mass index, kg/m
2

0.75
Underweight: <18.5 38 (4.7) 22 (69.7) 5 (6.3) 11 (24.0)

Normal: 18.5-24.9 468 (38.1) 296 (69.0) 82 (14.2) 90 (16.8)

Overweight: 25-29.9 371 (28.5) 242 (64.2) 52 (15.8) 77 (20.0)

Obese: ≥30 455 (28.7) 293 (64.8) 68 (14.2) 94 (21.0)

Age at menarche

0.06
≤11 392 (31.4) 258 (68.5) 52 (9.7) 82 (21.9)

12 405 (31.8) 244 (63.4) 71 (18.4) 90 (18.2)

13 291 (20.6) 184 (64.3) 45 (13.2) 62 (22.5)

≥14 244 (16.2) 167 (71.3) 39 (16.4) 38 (12.2)

Menopausal status

0.002
Premenopausal 1,181 (92.0) 774 (67.7) 185 (14.1) 222 (18.3)

Peri- and Post-menopausal 151 (8.0) 79 (52.6) 22 (16.8) 50 (30.6)

Joint parity & age (in years) at first
full-term pregnancy status

<0.001
Nulliparous 380 (38.2) 255 (73.8) 59 (12.1) 66 (14.2)

1-2, <25 294 (20.0) 177 (61.1) 35 (10.2) 82 (28.7)

1-2, ≥25 309 (21.4) 213 (65.8) 62 (21.7) 34 (12.5)

3+, <25 267 (15.4) 146 (52.6) 35 (13.4) 86 (33.9)

3+, ≥25 82 (5.0) 62 (77.9) 16 (18.7) 4 (3.5)

History of breast cancer among first
degree relative

0.23
No 1,155 (88.2) 748 (67.7) 182 (14.1) 225 (18.2)

Yes 111 (7.4) 69 (57.8) 14 (14.4) 28 (27.8)

Missing 66 (4.4) 36 (55.9) 11 (17.3) 19 (26.8)

Lifetime alcohol use status, g/day

<0.001
0 (Abstainers) 425 (30.3) 340 (81.5) 32 (5.5) 53 (13.0)

0.1-6.9 393 (30.3) 269 (68.8) 69 (15.1) 55 (16.2)

7-13.9 212 (15.7) 124 (64.7) 37 (15.7) 51 (19.6)

14-27.9 184 (14.8) 91 (57.6) 44 (19.1) 49 (23.3)

≥28 118 (8.9) 29 (25.4) 25 (30.8) 64 (43.8)

98
a
n=49 participants missing smoking status not included. All values are absolute frequencies and sample weighted row
percentages unless otherwise specified. Missing categories are not included in chi-square p-value estimates.
b
Absolute frequencies and sample weighted column percentages are presented for the total population
.

Abbreviations: CI, confidence interval; g, grams; kg, kilograms; m, meter.

99
Table 4.2. Smoking characteristics of women in the Young Women’s Health History Study by breast cancer status and subtype
(n=3193)
Breast Cancer Status Breast Cancer Subtype
b

Control Case Luminal A Luminal B HER2-type TNBC
N (W%)
a
1,332 (50.1%) 1,725 (49.9%) 666 (21.3%) 538 (15.2%) 102 (2.7%) 294 (7.4%)
Ever smoker status

Never smoker 853 (61.9) 1,076 (60.6) 404 (59.0) 353 (64.1) 50 (45.3) 193 (64.1)
Ever smoker 479 (38.2) 649 (39.4) 262 (41.0) 185 (36.0) 52 (54.7) 101 (36.0)
Personal smoking status

Never smoker 853 (61.9) 1,076 (60.6) 404 (59.0) 353 (64.1) 50 (45.3) 193 (64.1)
Former smoker 207 (17.9) 336 (21.2) 136 (22.7) 103 (19.9) 27 (27.9) 42 (15.7)
≥10 years since quitting
c
95 (10.7) 181 (12.2) 79 (13.6) 52 (11.1) 13 (15.0) 21 (8.1)
<10 years since quitting
c
112 (7.3) 153 (9.0) 57 (9.1) 50 (8.6) 14 (12.9) 20 (7.3)
Current smoker 272 (20.2) 313 (18.2) 126 (18.3) 82 (16.1) 25 (26.8) 59 (20.2)
Lifetime smoking pack-years

Never smoker 853 (61.9) 1,076 (60.6) 404 (59.0) 353 (64.1) 50 (45.3) 193 (64.1)
<5 226 (14.4) 250 (14.9) 92 (14.7) 81 (15.8) 18 (18.4) 37 (12.2)
5-19 186 (15.4) 286 (16.9) 123 (18.5) 80 (14.6) 21 (20.8) 40 (14.2)
≥20 67 (8.4) 107 (7.3) 47 (7.8) 21 (5.1) 12 (14.3) 22 (8.8)
Missing 0 (0.0) 6 (0.3) 0 (0.0) 3 (0.5) 1 (1.2) 2 (0.8)
Smoking initiation timing - first full-term
pregnancy
d

Never smoker 598 (62.5) 792 (61.2) 300 (58.9) 243 (63.8) 41 (48.8) 146 (67.2)
After FFTP 94 (7.3) 73 (5.0) 26 (4.3) 21 (5.3) 4 (5.2) 14 (5.9)
Before FFTP 259 (30.2) 394 (33.7) 170 (36.9) 108 (30.9) 33 (46.0) 52 (26.4)
Missing 1 (0.1) 2 (0.1) 0 (0.0) 0 (0.0) 0 (0.0) 1 (0.4)
Age smoking initiated, years

Never smoker 853 (61.9) 1,076 (60.6) 404 (59.0) 353 (64.1) 50 (45.3) 193 (64.1)
<18 233 (21.1) 370 (23.5) 154 (24.9) 102 (20.9) 33 (35.6) 56 (20.4)
18-24 207 (14.3) 204 (11.6) 81 (11.6) 62 (11.5) 14 (15.0) 29 (10.2)
≥25 39 (2.8) 74 (4.3) 27 (4.6) 20 (3.4) 5 (4.1) 16 (5.3)
Missing 0 (0.0) 1 (0.05) 0 (0.0) 1 (0.2) 0 (0.0) 0 (0.0)
Time since initiation, years

Never smoker 853 (61.9) 1,076 (60.6) 404 (59.0) 353 (64.1) 50 (45.3) 193 (64.1)
<20 201 (7.2) 152 (7.4) 47 (5.5) 47 (7.6) 12 (10.5) 33 (10.6)
20-29 187 (18.9) 328 (18.1) 139 (19.6) 101 (17.8) 21 (20.8) 43 (14.1)
≥30 91 (12.0) 168 (13.8) 76 (15.8) 36 (10.4) 19 (23.4) 25 (11.2)
Missing 0 (0.0) 1 (0.05) 0 (0.0) 1 (0.2) 0 (0.0) 0 (0.0)
Average number of cigarettes per day

Never smoker 853 (61.9) 1,076 (60.6) 404 (59.0) 353 (64.1) 50 (45.3) 193 (64.1)

100
Breast Cancer Status Breast Cancer Subtype
b

Control Case Luminal A Luminal B HER2-type TNBC
<5 125 (8.5) 161 (9.3) 58 (9.0) 53 (9.8) 9 (9.5) 26 (8.6)
5-19 236 (18.1) 301 (18.2) 128 (20.1) 87 (17.0) 24 (22.3) 39 (13.9)
≥20 118 (11.5) 184 (11.8) 76 (11.9) 44 (9.0) 18 (21.8) 35 (13.0)
Missing 0 (0.0) 3 (0.2) 0 (0.0) 1 (0.1) 1 (1.2) 1 (0.4)
a
n=136 participants missing smoking status not included. All values are absolute frequencies and sample weighted column percentages.
b
n=125 participants missing breast cancer subtype not included in analyses by breast cancer subtype
.

c
n=543 former smokers. Time since smoking cessation missing for n=2 former smokers.
d
Among n=2,213 parous women.
Abbreviations: FFTP, first full-term pregnancy; HER2, human epidermal growth factor receptor 2; TNBC, triple negative breast cancer.

101
Table 4.3. Multivariable-adjusted odds ratios and 95% CI
a
for the association of personal smoking history and risk of breast cancer
overall and by tumor subtype

Overall Luminal A Luminal B HER2-type TNBC

Adjusted OR (95% CI) Adjusted OR (95% CI) Adjusted OR (95% CI) Adjusted OR (95% CI) Adjusted OR (95% CI) Phet
Ever smoker status

0.02
Never smoker 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

Ever smoker 1.20 (0.99-1.46) 1.33 (1.06-1.68) 1.05 (0.78-1.42) 1.96 (1.22-3.15) 0.94 (0.68-1.28)

Personal smoking status

0.10
Never smoker 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

Former smoker 1.25 (0.98-1.61) 1.35 (1.01-1.81) 1.12 (0.75-1.66) 2.08 (1.24-3.48) 0.90 (0.61-1.33)

≥10 years since quitting
b
1.16 (0.83-1.63) 1.26 (0.85-1.88) 1.05 (0.67-1.67) 1.91 (1.03-3.56) 0.79 (0.44-1.45)

<10 years since quitting
b
1.37 (0.97-1.92) 1.52 (1.01-2.27) 1.26 (0.77-2.06) 2.29 (1.07-4.88) 0.88 (0.50-1.55)

Current smoker 1.14 (0.90-1.45) 1.31 (0.94-1.82) 0.97 (0.70-1.34) 1.81 (0.94-3.48) 0.97 (0.64-1.49)

Lifetime smoking pack-years

0.08
Never smoker 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

<5 1.12 (0.87-1.44) 1.15 (0.84-1.59) 1.08 (0.75-1.56) 1.79 (1.06-3.03) 0.83 (0.53-1.30)

5-19 1.33 (1.00-1.76) 1.59 (1.14-2.23) 1.11 (0.75-1.65) 1.92 (0.92-4.00) 0.89 (0.57-1.40)

≥20 1.09 (0.76-1.55) 1.24 (0.77-1.98) 0.78 (0.41-1.46) 2.26 (1.05-4.88) 1.13 (0.62-2.07)

Ptrend 0.11 0.02 0.85 0.02 0.96

Smoking initiation timing - first full-
term pregnancy
c

0.19
Never smoker 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

After FFTP 0.90 (0.59-1.37) 0.89 (0.49-1.59) 0.99 (0.52-1.90) 0.87 (0.25-3.07) 0.67 (0.33-1.36)

Before FFTP 1.25 (1.02-1.54) 1.43 (1.09-1.88) 1.16 (0.83-1.62) 1.74 (0.97-3.14) 0.83 (0.56-1.25)

Age smoking initiated, years

0.29
Never smoker 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

<18 1.33 (1.05-1.69) 1.51 (1.14-2.00) 1.16 (0.81-1.67) 2.35 (1.34-4.10) 0.96 (0.65-1.41)

18-24 0.91 (0.69-1.18) 0.95 (0.69-1.31) 0.83 (0.57-1.22) 1.39 (0.69-2.78) 0.72 (0.45-1.14)

≥25 1.91 (1.22-2.99) 2.22 (1.29-3.85) 1.45 (0.74-2.84) 2.18 (0.72-6.62) 1.96 (1.03-3.73)

Ptrend 0.17 0.07 0.94 0.04 0.94

Time since smoking initiated, years

0.14
Never smoker 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

<20 1.25 (0.93-1.66) 1.22 (0.83-1.79) 1.02 (0.67-1.57) 2.03 (1.07-3.87) 1.20 (0.76-1.91)

20-29 1.09 (0.85-1.39) 1.25 (0.93-1.67) 1.01 (0.72-1.42) 1.53 (0.85-2.74) 0.78 (0.52-1.17)

≥30 1.39 (1.02-1.91) 1.55 (1.07-2.24) 1.15 (0.67-1.96) 2.74 (1.29-5.79) 0.99 (0.56-1.74)

Ptrend 0.05 0.01 0.69 0.01 0.57

Average number of cigarettes per day

0.12
Never smoker 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

<5 1.21 (0.90-1.64) 1.24 (0.82-1.87) 1.17 (0.75-1.81) 1.59 (0.78-3.26) 1.02 (0.59-1.77)

5-19 1.17 (0.90-1.51) 1.37 (1.02-1.84) 1.05 (0.71-1.56) 1.69 (0.93-3.05) 0.75 (0.49-1.15)

102

Overall Luminal A Luminal B HER2-type TNBC

Adjusted OR (95% CI) Adjusted OR (95% CI) Adjusted OR (95% CI) Adjusted OR (95% CI) Adjusted OR (95% CI) Phet
≥20 1.23 (0.93-1.64) 1.35 (0.92-1.99) 0.93 (0.58-1.49) 2.60 (1.37-4.95) 1.13 (0.73-1.75)

Ptrend 0.09 0.03 0.96 0.01 0.82

a
Estimation of BC risk overall or by cancer subtype was relative to controls for all analyses. Adjusted for study site (Detroit, Los Angeles); age at diagnosis (continuous); HHP
(≥200%, <200% of FPL, missing); first degree family member w/ BC (Y, N, missing); BMI 12 months before reference date (underweight, normal, overweight, obese); alcohol use
(0, 0.1-6.9, 7-13.9, 14-27.9 and ≥28 grams/day); parity/age at FFTP (nulliparous, 1-2 children and <25 years, 1-2 children and ≥25, 3+ children and <25 years, 3+ children and
≥25); menopausal status (premenopausal, peri- & post-menopausal).
b
Among former smokers.
c
Among parous women.
Bold values are significant at P<0.05
Abbreviations: BC, breast cancer; BMI, body mass index; CI, confidence interval; ref, reference; FFTP, first full-term pregnancy; FPL, federal poverty level; HER2, human
epidermal growth factor receptor 2; het, heterogeneity; HHP, household poverty; N, no; OR, odds ratio; TNBC, triple negative breast cancer; Y, yes.

103
Table 4.4. Multivariable-adjusted odds ratios and 95% CI
a
for the association of personal smoking history and risk of breast cancer by
race and HHP

Non-Hispanic White Non-Hispanic Black

HHP ≥200% HHP <200%

Adjusted OR (95% CI) Adjusted OR (95% CI) Pint Adjusted OR (95% CI) Adjusted OR (95% CI) Pint
Ever smoker status

0.25

0.84
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

Ever smoker 1.37 (1.04-1.82) 1.01 (0.73-1.40)

1.27 (0.96-1.68) 1.10 (0.80-1.52)

Personal smoking status 0.14 0.03
c

Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)
Former smoker 1.34 (0.98-1.85) 1.34 (0.86-2.09) 1.17 (0.85-1.59) 1.73 (1.10-2.73)
≥10 years since quitting
b
1.28 (0.87-1.88) 0.98 (0.51-1.90) 1.11 (0.73-1.66) 1.65 (0.81-3.33)
<10 years since quitting
b
1.43 (0.94-2.18) 1.71 (0.93-3.14) 1.26 (0.83-1.91) 1.81 (1.04-3.15)
Current smoker 1.43 (1.00-2.04) 0.86 (0.58-1.30) 1.54 (1.02-2.32) 0.87 (0.60-1.27)
Lifetime smoking pack-years

0.10

0.99
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<5 1.18 (0.86-1.63) 1.02 (0.66-1.56)

1.16 (0.84-1.59) 1.06 (0.67-1.67)

5-19 1.71 (1.16-2.52) 0.86 (0.57-1.30)

1.42 (0.97-2.09) 1.28 (0.85-1.91)

≥20 1.18 (0.75-1.87) 1.53 (0.77-3.03)

1.21 (0.66-2.20) 0.88 (0.49-1.58)

Ptrend 0.04 0.83

0.11 0.90

Smoking initiation timing - first full-
term pregnancy
d

0.52

0.62
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

After FFTP 1.11 (0.41-2.97) 0.75 (0.48-1.17)

1.14 (0.53-2.48) 0.80 (0.46-1.38)

Before FFTP 1.44 (1.06-1.96) 1.13 (0.75-1.69)

1.25 (0.91-1.71) 1.24 (0.85-1.82)

Age smoking initiated, years

0.73

0.42
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<18 1.58 (1.16-2.16) 1.09 (0.72-1.66)

1.46 (1.02-2.08) 1.16 (0.78-1.74)

18-24 0.99 (0.67-1.44) 0.78 (0.53-1.16)

0.89 (0.61-1.30) 0.93 (0.61-1.42)

≥25 2.21 (0.96-5.08) 1.50 (0.81-2.77)

2.65 (1.19-5.92) 1.29 (0.65-2.55)

Ptrend 0.16 0.84

0.19 0.72

Time since smoking initiated, years

0.42

0.92
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<20 1.36 (0.89-2.08) 1.04 (0.66-1.63)

1.40 (0.88-2.22) 1.12 (0.71-1.79)

20-29 1.19 (0.85-1.66) 1.04 (0.71-1.52)

1.15 (0.80-1.65) 0.97 (0.67-1.40)

≥30 1.77 (1.18-2.65) 0.90 (0.50-1.60)

1.44 (0.92-2.26) 1.32 (0.74-2.34)

Ptrend 0.01 0.88

0.10 0.50

Average number of cigarettes per day

0.45

0.28
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<5 1.45 (0.94-2.24) 0.84 (0.54-1.31)

1.49 (0.98-2.27) 0.89 (0.51-1.53)

5-19 1.30 (0.90-1.88) 1.06 (0.68-1.63)

1.12 (0.76-1.64) 1.24 (0.85-1.83)

≥20 1.41 (0.96-2.05) 1.14 (0.63-2.04)

1.35 (0.85-2.15) 1.03 (0.65-1.63)

Ptrend 0.08 0.68

0.18 0.57

104

a
Estimation of BC risk was relative to controls for all analyses. Adjusted for study site (Detroit, Los Angeles); age at diagnosis (continuous); HHP (≥200%, <200% of FPL); first
degree family member w/ BC (Y, N, unknown); BMI 12 months before reference date (underweight, normal, overweight, obese); alcohol use (0, 0.1-6.9, 7-13.9, 14-42 and >42
grams/day); parity/age at FFTP (nulliparous, 1-2 children and <25 years, 1-2 children and 25-29 years, 1-2 children 30+ years, 3+ children and <25 years, 3+ children and 25-29
years, and 3+ children and 30+ years); menopausal status (premenopausal, peri- & post-menopausal). Models stratified by HHP are not adjusted for HHP.
b
Among former smokers.
c
Interaction term= HHP by personal smoking status (never, former, current).
c
Among parous women.
Bold values are significant at P<0.05 or Pint<0.10
Abbreviations: BC, breast cancer; BMI, body mass index; CI, confidence interval; FFTP, first full-term pregnancy; FPL, federal poverty level; HHP, household poverty; int,
interaction; N, no; OR, odds ratio; ref, reference; Y, yes.

105
Supplementary Table 4.1. Association of personal smoking history and risk of breast cancer by BC subtype and race
a

Non-Hispanic White Non-Hispanic Black Non-Hispanic White Non-Hispanic Black

Adjusted OR (95% CI) Adjusted OR (95% CI) Pint Adjusted OR (95% CI) Adjusted OR (95% CI) Pint

Luminal A Luminal B

Ever smoker status

0.44

0.88
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

Ever smoker 1.39 (1.03-1.89) 1.00 (0.61-1.66)

1.13 (0.74-1.71) 1.11 (0.69-1.77)

Personal smoking status

0.18

0.53
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

Former smoker 1.35 (0.95-1.93) 1.26 (0.67-2.37)

1.12 (0.70-1.81) 1.44 (0.73-2.85)

Current smoker 1.48 (0.99-2.21) 0.88 (0.47-1.62)

1.13 (0.69-1.87) 0.94 (0.53-1.68)

Smoking pack-years

0.37

0.63
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<5 1.16 (0.79-1.69) 1.02 (0.55-1.88)

1.12 (0.70-1.79) 1.14 (0.62-2.11)

5-19 1.85 (1.18-2.90) 0.88 (0.46-1.69)

1.34 (0.81-2.20) 0.99 (0.53-1.85)

≥20 1.16 (0.65-2.09) 1.52 (0.60-3.87)

0.75 (0.31-1.82) 1.38 (0.44-4.36)

p-trend 0.048 0.85

0.88 0.74

Smoking initiation timing - first full-
term pregnancy
d

0.85

0.77
Never smoker 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

After 1st FTP 0.89 (0.27-2.97) 0.86 (0.43-1.70) 1.63 (0.30-8.80) 0.88 (0.43-1.81)

Before 1st FTP 1.44 (1.00-2.08) 1.28 (0.64-2.53) 1.28 (0.79-2.05) 1.23 (0.66-2.29)

Age smoking initiated, years

0.24

0.99
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<18 1.71 (1.20-2.43) 0.83 (0.43-1.59)

1.29 (0.81-2.06) 1.19 (0.58-2.44)

18-24 0.85 (0.55-1.31) 1.02 (0.57-1.83)

0.85 (0.49-1.47) 0.91 (0.54-1.54)

≥25 2.74 (1.12-6.72) 1.33 (0.59-2.98)

1.87 (0.49-7.17) 1.47 (0.58-3.75)

p-trend 0.19 0.68

0.79 0.71

Time since smoking initiated, years

0.77

0.94
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<20 1.28 (0.76-2.16) 1.05 (0.49-2.27)

1.02 (0.54-1.92) 1.20 (0.65-2.22)

20-29 1.26 (0.85-1.87) 1.04 (0.57-1.92)

1.11 (0.70-1.76) 0.95 (0.51-1.77)

≥30 1.73 (1.11-2.69) 0.89 (0.40-2.00)

1.26 (0.63-2.51) 1.20 (0.47-3.03)

p-trend 0.02 0.90

0.50 0.79

Average number of cigarettes per day

0.91

0.67
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<5 1.31 (0.77-2.24) 0.95 (0.48-1.86)

1.34 (0.76-2.37) 0.96 (0.46-2.00)

5-19 1.42 (0.95-2.12) 1.03 (0.54-1.95)

1.10 (0.65-1.87) 1.17 (0.68-2.01)

≥20 1.41 (0.87-2.28) 1.03 (0.41-2.59)

0.98 (0.51-1.85) 1.19 (0.54-2.64)

p-trend 0.0499 0.93

0.85 0.53

106
Supplementary Table 4.1 (continued). Association of personal smoking history and risk of breast cancer by BC subtype and race
a

Non-Hispanic White Non-Hispanic Black Non-Hispanic White Non-Hispanic Black

Adjusted OR (95% CI) Adjusted OR (95% CI) Pint Adjusted OR (95% CI) Adjusted OR (95% CI) Pint

HER2-Type
b
TNBC
b

Ever smoker status

0.11

0.47
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

Ever smoker 2.87 (1.62-5.10) 1.16 (0.48-2.81)

1.02 (0.65-1.63) 0.93 (0.57-1.53)

Personal smoking status

0.26

0.67
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

Former smoker 2.73 (1.46-5.12) 2.09 (0.68-6.45)

1.03 (0.62-1.70) 0.74 (0.38-1.42)

Current smoker 3.17 (1.44-7.00) 0.78 (0.25-2.44)

1.02 (0.52-1.99) 1.03 (0.57-1.87)

Smoking pack-years

0.19

0.25
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<5 2.48 (1.30-4.73) 0.93 (0.27-3.22)

0.75 (0.40-1.42) 0.92 (0.50-1.68)

5-19 3.24 (1.34-7.85) 0.98 (0.30-3.21)

1.23 (0.66-2.30) 0.66 (0.32-1.36)

≥20 2.91 (1.05-8.11) 2.97 (0.63-13.96)

1.17 (0.57-2.42) 2.21 (0.79-6.22)

p-trend 0.003 0.41

0.60 0.99

Smoking initiation timing - first full-
term pregnancy
c

0.41

0.73
d

Never smoker 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

After 1st FTP 1.99 (0.14-27.57) 0.49 (0.13-1.89) --- 0.75 (0.39-1.46)

Before 1st FTP 2.70 (1.22-5.98) 1.15 (0.42-3.10) 0.89 (0.48-1.68) 0.99 (0.53-1.85)

Age smoking initiated, years

0.40

0.54
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<18 3.60 (1.84-7.03) 1.53 (0.47-4.90)

1.14 (0.67-1.93) 1.21 (0.64-2.28)

18-24 2.01 (0.82-4.96) 0.62 (0.18-2.13)

0.83 (0.44-1.57) 0.48 (0.21-1.07)

≥25 1.57 (0.15-16.85) 2.56 (0.55-11.96)

1.40 (0.32-6.20) 1.85 (0.83-4.14)

p-trend 0.02 0.72

0.94 0.91

Time since smoking initiated, years

0.30

0.02
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<20 2.80 (1.19-6.59) 1.13 (0.32-3.98)

1.30 (0.61-2.76) 1.04 (0.55-1.95)

20-29 2.01 (1.01-3.98) 1.31 (0.49-3.54)

0.68 (0.39-1.20) 1.08 (0.61-1.89)

≥30 5.17 (1.98-13.48) 0.97 (0.22-4.23)

1.53 (0.76-3.07) 0.41 (0.16-1.05)

p-trend 0.001 0.85

0.73 0.39

Average number of cigarettes per day

0.29

0.92
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<5 2.55 (1.01-6.43) 0.54 (0.12-2.53)

1.05 (0.43-2.58) 0.86 (0.43-1.72)

5-19 2.42 (1.22-4.79) 1.21 (0.37-3.98)

0.79 (0.42-1.47) 0.78 (0.40-1.53)

≥20 3.67 (1.59-8.46) 2.00 (0.62-6.46)

1.38 (0.80-2.38) 1.28 (0.49-3.35)

p-trend 0.001 0.33

0.62 0.89

a
Multivariable odds ratios and 95% CIs are presented. Adjusted for study site (Detroit, Los Angeles); age at diagnosis; HHP (≥200%, <200% of FPL); first degree family member
w/ BC (Y, N, unknown); BMI 12 months before reference date (underweight, normal, overweight, obese); alcohol use (0, 0.1-6.9, 7-13.9, 14-42 and >42 grams/day); age at first
birth/parity (nulliparous, 1-2 children and <25 years, 1-2 children and 25-29 years, 1-2 children 30+ years, 3+ children and <25 years, 3+ children and 25-29 years, and 3+ children
and 30+ years); premenopausal status (Y, N). Models stratified by race are not adjusted for race.

107
b
Models were adjusted for a new BMI 12 months before reference date term with the first two categories (underweight and normal) combined because of zero cell count in the
analysis by breast cancer subtype.
c
Among parous women.
d
The interaction p-value is estimated with for race by initiation before first full-term pregnancy only because of zero count in cell for initiation after first FTP among NH Whites.
Abbreviations: BC, breast cancer; BMI, body mass index; CI, confidence interval; FPL, federal poverty level; FTP, full-term pregnancy; HER2, human epidermal growth factor
receptor 2; HHP, household poverty; int, interaction; N, no; NH, non-Hispanic; OR, odds ratio; ref, reference; TNBC, triple negative breast cancer; Y, yes.

108
Supplementary Table 4.2. Association of personal smoking history and risk of breast cancer by BC subtype and HHP

HHP ≥200% HHP <200% HHP ≥200% HHP <200%

Adjusted OR (95% CI) Adjusted OR (95% CI) Pint Adjusted OR (95% CI) Adjusted OR (95% CI) Pint

Luminal A Luminal B

Ever smoker status

0.19

0.75
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

Ever smoker 1.25 (0.90-1.74) 1.50 (0.95-2.36)

1.11 (0.73-1.68) 0.97 (0.55-1.73)

Personal smoking status

0.08

0.05
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

Former smoker 1.18 (0.81-1.71) 2.26 (1.30-3.92)

1.04 (0.64-1.67) 1.53 (0.78-3.01)

Current smoker 1.43 (0.90-2.28) 1.21 (0.69-2.11)

1.30 (0.76-2.20) 0.74 (0.41-1.34)

Smoking pack-years

0.63

0.89
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<5 1.05 (0.69-1.58) 1.38 (0.74-2.59)

1.16 (0.73-1.83) 1.03 (0.48-2.19)

5-19 1.55 (0.98-2.45) 1.75 (0.98-3.15)

1.12 (0.65-1.94) 1.23 (0.63-2.40)

≥20 1.12 (0.55-2.25) 1.31 (0.57-3.00)

0.85 (0.29-2.48) 0.54 (0.20-1.50)

p-trend 0.14 0.21

0.87 0.48

Smoking initiation timing - first full-
term pregnancy
d

0.17

0.98
Never smoker 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

After 1st FTP 1.01 (0.39-2.62) 0.84 (0.35-1.98) 1.22 (0.41-3.64) 1.06 (0.42-2.69)

Before 1st FTP 1.21 (0.79-1.84) 2.00 (1.17-3.43) 1.16 (0.70-1.90) 1.15 (0.60-2.22)

Age smoking initiated, years

0.10

0.67
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<18 1.34 (0.89-2.03) 1.80 (1.04-3.10)

1.35 (0.82-2.25) 0.87 (0.45-1.69)

18-24 0.88 (0.55-1.41) 1.12 (0.64-1.96)

0.80 (0.49-1.31) 0.94 (0.44-2.02)

≥25 3.48 (1.48-8.21) 0.83 (0.21-3.28)

1.49 (0.46-4.87) 1.70 (0.60-4.85)

p-trend 0.14 0.58

0.99 0.70

Time since smoking initiated, years

0.30

0.79
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<20 1.34 (0.74-2.46) 1.13 (0.61-2.08)

1.04 (0.59-1.83) 1.06 (0.47-2.39)

20-29 1.25 (0.82-1.91) 1.17 (0.70-1.94)

1.00 (0.61-1.65) 1.03 (0.56-1.90)

≥30 1.19 (0.73-1.96) 2.44 (1.12-5.32)

1.43 (0.69-2.95) 0.74 (0.28-1.94)

p-trend 0.24 0.045

0.49 0.72

Average number of cigarettes per day

0.42

0.96
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<5 1.28 (0.75-2.20) 1.14 (0.54-2.44)

1.37 (0.78-2.38) 1.04 (0.43-2.53)

5-19 1.17 (0.75-1.82) 1.75 (0.98-3.10)

1.08 (0.62-1.86) 1.04 (0.54-2.01)

≥20 1.38 (0.81-2.35) 1.41 (0.74-2.69)

0.90 (0.42-1.95) 0.83 (0.38-1.85)

p-trend 0.21 0.10

0.96 0.74

109
Supplementary Table 4.2. (continued). Association of personal smoking history and risk of breast cancer by BC subtype and SEP

HHP ≥200% HHP <200% HHP ≥200% HHP <200%

Adjusted OR (95% CI) Adjusted OR (95% CI) Pint Adjusted OR (95% CI) Adjusted OR (95% CI) Pint

HER2-Type
b
TNBC
b

Ever smoker status

0.15

0.26
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

Ever smoker 2.67 (1.41-5.06) 1.30 (0.62-2.71)

0.99 (0.65-1.52) 0.79 (0.47-1.32)

Personal smoking status

0.39

0.05
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

Former smoker 2.77 (1.40-5.47) 1.27 (0.43-3.76)

0.78 (0.48-1.28) 1.19 (0.53-2.64)

Current smoker 2.44 (0.99-5.98) 1.31 (0.56-3.10)

1.41 (0.76-2.62) 0.64 (0.34-1.20)

Smoking pack-years

0.36

0.45
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<5 2.34 (1.12-4.87) 0.98 (0.34-2.86)

0.90 (0.53-1.54) 0.56 (0.24-1.30)

5-19 2.33 (0.87-6.26) 1.94 (0.77-4.91)

1.10 (0.60-2.01) 0.76 (0.40-1.44)

≥20 4.31 (1.50-12.40) 0.95 (0.23-3.86)

0.90 (0.33-2.41) 1.00 (0.37-2.70)

p-trend 0.003 0.61

0.99 0.73

Smoking initiation timing - first full-
term pregnancy
c

0.68

0.65
Never smoker 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

After 1st FTP 0.41 (0.03-5.62) 1.12 (0.23-5.38) 0.76 (0.25-2.32) 0.56 (0.22-1.44)

Before 1st FTP 1.87 (0.79-4.43) 1.43 (0.59-3.47) 0.85 (0.46-1.54) 0.69 (0.37-1.29)

Age smoking initiated, years

0.04

0.58
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<18 4.10 (2.04-8.24) 1.01 (0.43-2.37)

1.12 (0.65-1.92) 0.82 (0.43-1.57)

18-24 0.94 (0.28-3.14) 1.96 (0.74-5.20)

0.76 (0.42-1.40) 0.41 (0.15-1.15)

≥25 4.09 (0.79-21.13) 0.63 (0.05-7.62)

1.70 (0.57-5.04) 1.99 (0.76-5.20)

p-trend 0.08 0.35

0.93 0.74

Time since smoking initiated, years

0.31

0.70
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<20 2.23 (0.83-5.98) 1.73 (0.61-4.96)

1.26 (0.63-2.54) 1.22 (0.63-2.38)

20-29 1.84 (0.74-4.56) 1.24 (0.54-2.84)

0.82 (0.48-1.39) 0.58 (0.29-1.13)

≥30 4.90 (1.86-12.93) 1.09 (0.29-4.10)

1.13 (0.49-2.62) 0.77 (0.24-2.50)

p-trend 0.002 0.71

0.96 0.30

Average number of cigarettes per day

0.15

0.44
Never smoker 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<5 2.85 (1.19-6.80) 0.26 (0.02-2.65)

1.21 (0.60-2.46) 0.69 (0.27-1.78)

5-19 1.84 (0.74-4.56) 1.73 (0.81-3.71)

0.86 (0.48-1.53) 0.56 (0.27-1.15)

≥20 4.08 (1.69-9.83) 1.41 (0.42-4.70)

1.07 (0.58-1.99) 1.05 (0.48-2.27)

p-trend 0.01 0.33

0.91 0.61

a
Multivariable odds ratios and 95% CIs are presented. Adjusted for study site (Detroit, Los Angeles); age at diagnosis; household poverty (≥200%, <200% of FPL); first degree
family member w/ BC (Y, N, unknown); BMI 12 months before reference date (underweight, normal, overweight, obese); alcohol use (0, 0.1-6.9, 7-13.9, 14-42 and >42

110
grams/day); age at first birth/parity (nulliparous, 1-2 children and <25 years, 1-2 children and 25-29 years, 1-2 children 30+ years, 3+ children and <25 years, 3+ children and 25-
29 years, and 3+ children and 30+ years); premenopausal status (Y, N). Models stratified by race are not adjusted for race. Models stratified by SEP are not adjusted for SEP.
b
Models were adjusted for a new BMI 12 months before reference date term with the first two categories (underweight and normal) combined because of zero cell counts in the
analysis by breast cancer subtype.
c
Among parous women.
Abbreviations: BC, breast cancer; BMI, body mass index; CI, confidence interval; FPL, federal poverty level; FTP, full-term pregnancy; HER2, human epidermal growth factor
receptor 2; HHP, household poverty; int, interaction; N, no; OR, odds ratio; ref, reference; TNBC, triple negative breast cancer; Y, yes.

111
Supplementary Table 4.3. Sensitivity analyses estimating the association between ever smoking and BC risk in listed subpopulations
Subpopulation N Overall Luminal A Luminal B HER2-type TNBC

Controls Cases
Adjusted OR
(95% CI)
Adjusted OR
(95% CI)
Adjusted OR
(95% CI)
Adjusted OR
(95% CI)
Adjusted OR
(95% CI)
P
het

Premenopausal
b
1181 1492
1.21
(0.97-1.49)
1.35
(1.05-1.74)
1.02
(0.73-1.40)
1.81
(1.10-3.00)
1.00
(0.71-1.41)
0.03
No first-degree family
history of BC
c

1155 1309
1.16
(0.93-1.44)
1.24
(0.95-1.62)
1.06
(0.76-1.48)
1.86
(1.15-3.01)
0.87
(0.62-1.21)
0.03
Alcohol abstainers
d
425 504
1.12
(0.93-1.35)
1.26
(0.71-2.23)
1.02
(0.56-1.86)
1.70
(0.46-6.27)
0.88
(0.43-1.77)
0.75
a
Multivariable odds ratios and 95% CIs are presented. Adjusted for study site (Detroit, Los Angeles); age at diagnosis; household poverty (≥200%, <200% of FPL); first degree
family member w/ BC (Y, N, unknown); BMI 12 months before reference date (underweight, normal, overweight, obese); alcohol use (0, 0.1-6.9, 7-13.9, 14-42 and >42
grams/day); age at first birth/parity (nulliparous, 1-2 children and <25 years, 1-2 children and 25-29 years, 1-2 children 30+ years, 3+ children and <25 years, 3+ children and 25-
29 years, and 3+ children and 30+ years); premenopausal status (Y, N). Models stratified by race are not adjusted for race. Models stratified by SEP are not adjusted for SEP.
b
Models were not adjusted for premenopausal status.
c
Models were not adjusted for first degree family history of BC.
d
Models were not adjusted for alcohol use.
Abbreviations: BC, breast cancer; BMI, body mass index; CI, confidence interval; FPL, federal poverty level; FTP, full-term pregnancy; HER2, human epidermal growth factor
receptor 2; HHP, household poverty; int, interaction; N, no; OR, odds ratio; ref, reference; TNBC, triple negative breast cancer; Y, yes.

112
CHAPTER FIVE: Lifetime composite smoking and risk for breast cancer subtypes among
Non-Hispanic Black and White women in the Young Women’s Health History Study

Authors: Ugonna Ihenacho
1
, Ann S. Hamilton
1
, Wendy J. Mack
1
, Anna H. Wu
1
, Jennifer B.
Unger
1
, Dorothy R. Pathak
2
, Richard T. Houang
3
, Michael F. Press
4
, Kendra L. Schwartz
5
, Lydia
R. Marcus
6
, Ellen M. Velie
6,7*

Institutional affiliations:
8. Department of Preventive Medicine, Keck School of Medicine, University of Southern
California, Los Angeles, CA, USA
9. Department of Epidemiology and Biostatistics, College of Human Medicine, Michigan State
University, East Lansing, MI, USA
10. Center for the Study of Curriculum, College of Education, Michigan State University, East
Lansing, MI, USA
11. Department of Pathology, Keck School of Medicine, University of Southern California, Los
Angeles, CA, USA
12. Department of Family Medicine and Public Health Sciences, School of Medicine, Karmanos
Cancer Institute, Wayne State University, Detroit, MI, USA
13. Joseph J. Zilber School of Public Health, University of Wisconsin-Milwaukee, Milwaukee,
WI, USA
14. Departments of Medicine and Pathology, Medical College of Wisconsin, Milwaukee, WI,
USA

113
5.1 Abstract
Background: Studies of lifetime active and secondhand smoke (SHS) exposures with breast
cancer (BC) risk have predominately been conducted among populations of White, educated
women. Few studies have been conducted in populations of racially and socioeconomically
diverse women to confirm the association and to explore lifetime smoking as a factor
contributing to racial and socioeconomic disparities in BC risk.
Methods: We evaluated the association of lifetime active smoke and SHS characteristics with
BC risk using data from the Young Women’s Health History (YWHHS) study. In total 1,812
cases and 1,381 controls were enrolled in this population-based case-control study of non-
Hispanic Black and White women 20-49 years of age. Sample-weighted logistic regression
analysis was used to evaluate the association overall with a polytomous regression for estimates
by BC subtype. We explored potential effect modification by prenatal smoke exposure status,
race and socioeconomic position (SEP).
Results: About 75% of the study population had SHS exposure in their lifetime. A 69%
decreased risk of human epidermal receptor 2 (HER2)-type BC was observed with ≥15 years
SHS exposure in childhood. A non-significant increased risk of TNBC was identified among
women with childhood SHS exposure. Lifetime composite exposure was a weak indicator of
breast cancer risk.
Conclusion: Lifetime composite smoke was not associated with BC risk in the YWHHS. Further
analysis is needed in populations with a larger sample of HER2-type and triple negative BC
subtypes to explore suggested findings.

5.2 Introduction

114
Breast cancer (BC) is the most commonly diagnosed cancer among women under age 50
years in the United States (US) excluding skin cancers (1). Although results have not been all
consistent, a recent pooled analysis of 14 cohort studies found that current smoking was
associated with a 7% increased risk of BC (2). However, the study population of this pooled
analysis was more than 90% non-Hispanic White (NHW), about 60% peri- and postmenopausal,
and educated (2). Therefore, additional studies are needed to examine the association between
smoking and BC among women under age 50 and to consider potential differences by race and
SEP.
Breast cancer risk varies by age, race and socioeconomic position (SEP). Black women
under age 40 years have a higher incidence of BC and higher BC mortality rate than White
women of the same age group (3). In 1998, Liu et.al. described BCs diagnosed between 1972 and
1992 in Los Angeles (LA) County and found that census tract socioeconomic status was
positively associated with risk of BC across all race groups and 10-year age groups (4). More
recently, the Black Women’s Health Study, a cohort of 52,425 Black women, found that
individual SEP (based on education) and neighborhood SEP measures were also positively
associated with BC risk, but this was only observed for estrogen receptor (ER)+ and not for ER-
BCs (5). In general, ER- tumors are commonly diagnosed among women of Black and Hispanic
race/ethnicity, low SEP, premenopausal BC cases, and smokers (6).
Further studies on the association between smoking and BC by race and SEP are
warranted due to differences in smoking behaviors between White and Black women, and
between women of low- and high-SEP. The Centers for Disease Control and Prevention (CDC)
reports that Black smokers tend to initiate smoking at a later age, have a longer smoking history,
and smoke fewer cigarettes per day compared to White smokers (7; 8). Similarly, individuals

115
living below the poverty level in the US are more likely to smoke more heavily compared to
those above the poverty level (7).
In the pooled analysis, the authors cautioned that they were unable to account for
secondhand smoke (SHS) exposures, which may bias their results toward the null (2). This
limitation warrants further exploration considering that the National Health and Nutrition
Examination Survey (NHANES 2013-2014) found that 20% of non-smoking adults in the US
have detectable serum cotinine levels, indicating recent SHS exposure (9). NHANES also
described that the SHS exposure prevalence among non-smokers was highest among non-
Hispanic Blacks (NHB; 50.3%), adults living in poverty (47.9%), and children aged 3-11 years
(37.9%) (9). Other studies have evaluated the association between active smoking and BC with
consideration of SHS exposure history in childhood and adulthood; however, many were
conducted in the US and Europe, and like the pooled analysis, included predominately White
populations and older, postmenopausal women (10-15).
In this study, we build on the current literature by characterizing active and lifetime SHS
exposures using a lifetime composite smoke exposure variable to estimate the association with
BC risk, overall and by BC subtype. Also, we use a racially and socioeconomically diverse
population of women under age 50 years to evaluate whether active smoking-related risk
association is modified by prenatal smoking exposure history, race or SEP.

5.3 Methods
5.3.1 Study population of cases and controls
Data used for this analysis are from a population-based case-control study of BC among
young women under 50 years of age, the Young Women’s Health History Study (YWHHS).

116
Participants were identified from residents of the LA County and Metropolitan Detroit
Surveillance, Epidemiology, and End Results (SEER) Registry areas. Case participants were
identified among women with histologically confirmed incident invasive primary BC diagnosed
between 2010 and 2015. Women diagnosed with Paget’s disease, breast lymphoma,
mesenchymal tumors, including sarcomas and hemangiosarcomas of the breast were excluded.
Controls were identified through area-based sampling methods from over 24,000 postal
addresses based on the 2010 US Census and were frequency matched to cases on race, five-year
age group and study region. In total, 1,812 women with invasive BC (1,130 NHW, 682 NHB;
59.8% case response rate) and 1,381 control participants (716 NHW, 665 NHB; 53% control
response rate) completed an in-person study interview.
Additional criteria for inclusion for all participants included: female, aged 20-49 years at
reference date, self-identified race/ethnicity as NHB or NHW, no previous diagnosis of any
cancer before the reference date except for cervical cancer in situ or non-melanoma skin cancer,
residing in LA County or the tri county (Macomb, Oakland and Wayne) metropolitan Detroit
area, born in the US, able to complete the study interview in English, physically and mentally
able to complete the interview, and not institutionalized. For cases, the study reference date was
defined as the date of histologically confirmed BC diagnosis. For controls, the reference date was
the date of the screening interview minus four months.
Institutional review boards at the University of Southern California, the University of
Wisconsin – Milwaukee, Michigan State University, Wayne State University, the Michigan
Department of Community Health, the Medical College of Wisconsin, the California Committee
for the Protection of Human Subjects, and the California Cancer Registry approved the study.

117
Written informed consent was obtained from each participant before study interviews were
conducted.

5.3.2 Tumor Subtyping
Tumor characteristics were obtained from the local SEER registry for BC cases included
in this study. Progesterone receptor (PR) status, ER status, human epidermal growth factor
receptor 2 (HER2) status, and tumor grade were evaluated to determine tumor subtyping.
Molecular subtypes were defined as follows: Luminal A (ER/PR+, HER2-, grade 1/2); Luminal
B (ER/PR+, HER2+ or HER2-, grade 3+); HER2-type (ER-, PR-, HER2+); and Triple Negative
BC (TNBC; ER-, PR- HER2-) (see Section 3.6.1) (16).

5.3.3 Lifestyle factors and tobacco exposures
Structured in-person interviews were conducted with study participants.
Sociodemographic information was collected to confirm study eligibility and on factors of
interest for the study. These characteristics included place of birth, age, race/ethnicity, duration
of residence in LA County or metropolitan Detroit, education, and household income 12 months
before the reference date and the number of household members supported by that income.
Characteristics pertaining to established and suspected BC risk factors included age at menarche,
pregnancy history, anthropometric measures in childhood, adolescence and adulthood, family
history of BC among first degree relatives, and menopausal status. Lifetime smoking histories
were collected, which included information on smoking status, SHS exposures before and after
age 18 years, and prenatal smoke exposures (in utero) (see Section 3.6.2).

118
5.3.4 Statistical analysis
The sociodemographic variables collected include self-reported race (NHB, NHW), age
at diagnosis/reference date, region (metropolitan Detroit, LA County), education (completed
high school or less; vocational school, associate’s degree, or some college; bachelor’s degree or
higher), and the household poverty level (HHP) as a measure of SEP. HHP was derived from
self-reported household income 12 months before reference date and the number of household
members supported by that income to calculate the household poverty percent of federal poverty
level (FPL) (≥200% of FPL, <200% of FPL). The values were calculated using the 2009-2014
US poverty thresholds (17). We chose the cut point of 200% of FPL to include as low SEP a
broader definition for this analysis than the standards of 133-200% that is typically used to
qualify for federal assistance programs (17; 18).
Established and suspected BC risk factors were characterized as: age at menarche (≤11,
12, 13, ≥14 years), menopausal status (premenopausal, peri- & postmenopausal), first degree
family history of BC (yes, no, missing/don’t know), body mass index (BMI) 12 months before
reference date (underweight: <18.5, normal: 18.5-24.9, overweight: 25.0-29.9, obese: ≥30
kg/m
2
), lifetime cumulative alcohol use (0, 0.1-6.9, 7-13.9, 14-27.9 and ≥28 grams/day), and a
joint variable of parity and age at first full-term pregnancy (FFTP) (nulliparous; 1-2 children and
<25 years; 1-2 children and ≥25 years; 3+ children and <25 years; 3+ children and ≥25 years)
(see Section 3.6.3).
Prenatal smoke exposure was characterized by the participant’s report of their biological
mother smoking while pregnant with them (no, yes, unknown). If the participants’ primary
childhood caregiver completed the caregiver’s survey and provided a yes/no response for
prenatal smoke exposure, then the caregiver’s response was substituted for don’t know/missing

119
responses among participants (n=186). Lifetime SHS exposure was defined as having a person
who smoked at least one cigarette a day in the participant’s presence for 6 months or more in
childhood (before age 18 years) or adulthood (after age 18 years up to 12 months before
reference date). SHS exposure is also described by duration of exposure in lifetime (<15 years
SHS exposure, ≥15 years SHS exposure), in childhood (<15 years SHS exposure, ≥15 years SHS
exposure), and in adulthood (<10 years SHS exposure, ≥10 years SHS exposure) based on the
median duration of exposure among controls in the study. Lifetime composite smoking status
was characterized by combinations of childhood and adulthood SHS exposure among non-
smokers and a category capturing personal smoking status (no personal or secondhand exposure;
childhood secondhand exposure only; adulthood secondhand exposure only; both childhood &
adulthood secondhand exposures; ever smoker). Ever active smokers were defined as
participants who smoked at least one cigarette a day for six months or more in their lifetime,
regardless of SHS status.
Matching characteristics and lifestyle factors were described by categories of lifetime
composite smoking status using a sample-weighted chi-square test for categorical variables and
sample-weighted Wald’s test for continuous variables. The relative risk of BC was estimated by
odds ratio using sample-weighted logistic regression models assessed for BC overall and by BC
subtype. Crude and multivariable adjusted models were designed and the adjusted odds ratios
(aOR) and 95% confidence intervals (CIs) for the association between lifetime cigarette smoking
characteristics and BC risk are presented. Multivariable polytomous logistic regression were
employed for estimating associations by BC subtype. Sample weights take into account the
sampling design and non-response among sampled households. Control status was the reference
outcome for analyses by BC status and BC subtype.

120
Three models were utilized to evaluate the associations of interest by BC status overall
and by BC subtype. First crude and multivariable adjusted models were developed to estimate
the association between lifetime composite smoking exposure and BC risk using the total study
population. Then two models were developed to estimate the association between the SHS
exposures and BC risk among a subpopulation of never smokers (n=1,929): (i) estimating the
association between childhood SHS exposure status (no, yes) and adulthood SHS exposure status
(no, yes) (mutually adjusted) and BC risk among never smokers; (ii) estimating the mutually
adjusted associations of duration of childhood SHS exposure (no, <15 years, ≥15 years SHS
exposure) and duration of adulthood SHS exposure (no, <10 years, ≥10 years SHS exposure)
with BC risk among never smokers.
Models were adjusted for region, age at diagnosis/reference date, first degree family
member with BC, age at menarche, BMI 12 months before reference date, parity/age at FFTP,
alcohol use, menopausal status, and HHP based on assessments for confounding in the
association between lifetime composite smoking and BC (overall and by subtype) or based on a
priori knowledge. Heterogeneity in the odds ratio estimates for polytomous regression models
were assessed by Wald’s tests. We investigated potential effect modification by prenatal smoke
exposure status, race and HHP, conducted stratified models, added cross-product interaction
terms of the smoking exposures by each stratum, and evaluated statistical interactions by Wald’s
test.
Personal smoking status (ever or never smoker) was missing for 136 participants (87
cases and 49 controls) who were excluded from all analyses. Missing data was captured as a
separate category for all other smoking characteristics but did not contribute to the estimated p-
values. Cases missing a specification for BC subtype were excluded from analyses by BC

121
subtype (n=125). Values were imputed for missing data for all covariates except first degree
family member with BC, where missing or unknown (n=38) was retained as a separate category.
Missing values for the covariates age at menarche (n=76), BMI 12 months before reference date
(n=9), alcohol use (n=20), and the joint indicator for parity/age at FFTP (n=1), were imputed
based on the sample median for race, HHP, and joint parity/age at FFTP. The study population
median values were used to impute missing values if the participant was also missing
information for HHP or joint parity/age at FFTP.
Tests were assessed at a significance level of P<0.05. All analyses were conducted using
Stata version 14.2 (StataCorp LLC, College Station, TX).

5.4 Results
Table 5.1 describes demographic and BC risk factors by lifetime composite smoke
exposure status among controls. Composite smoke exposure differed by region with a higher
proportion of women with no personal or SHS exposure in LA County (36.2%) compared to
Metropolitan Detroit (23.8%) (p=0.001). Women with no personal or SHS exposure were
younger (mean age of 30.5 years) than women with both childhood and adulthood SHS
exposures (mean age of 36.9 years) and ever smokers (mean age of 36.1) (p<0.001). Control
participants with a bachelor’s degree or higher had the highest proportion of no personal or SHS
exposure (39.5%) and participants with a high school diploma or less had the highest proportion
of ever smokers (43.1%) (p<0.001). Control participants with no personal or SHS exposure were
more likely to be premenopausal (30.0%) whereas ever smokers were more likely to be peri-
/postmenopausal (47.4%). There was a statistically significant difference in joint parity/age at
FFTP by composite smoking status (p=0.007). There was a higher proportion of women with no

122
personal or SHS exposure among alcohol abstainers (34.7%) and a higher proportion of ever
smokers among participants that consumed ≥ 28 grams/day (74.6%) compared the other
cumulative alcohol consumption categories (p<0.001). No significant associations were observed
among the other sample characteristics assessed including family history of breast cancer, adult
BMI, and age at menarche.
We describe the smoking characteristics of the controls (n=1,332) and cases (n=1,725) in
the study and by BC subtype (Table 5.2). Prenatal smoke exposures were similar between cases
and controls (25.1% vs. 24.7%, respectively) with a slightly higher proportion of control
participants responding with an unknown prenatal smoke exposure status. The highest proportion
of prenatal smoking exposure was observed among women diagnosed with the Luminal B BC
subtype (26.5%). About 78% of controls and 76% of cases reported any SHS exposure in their
lifetime. Lifetime SHS exposures among breast cancer cases ranged between 73% for Luminal B
to 81% for TNBC. About 50% of the study population had ≥15 years of lifetime SHS exposure.
About 65% of the study population was exposed to SHS in childhood; slightly more than half of
the participants with childhood SHS exposure were exposed for ≥15 years in childhood. The
highest proportion of childhood SHS exposures ≥15 years was observed among cases diagnosed
with TNBC (41.7%). A little more than half of the study population had SHS exposures in
adulthood (55.9% among controls vs. 53.2% among cases). Cases with HER2-type BC had the
highest proportion of ≥10 years of SHS exposure in adulthood (39.1%). The distribution of
composite smoking exposure was similar by BC status.
We did not identify any significant associations between lifetime composite smoke
exposure and BC risk overall or by subtype (Table 5.3). An increased risk of HER2-type BC is
suggested with ever smoking, but the odds ratio did not reach statistical significance (aOR 1.46;

123
95% CI 0.81, 2.62). Risk of TNBC was elevated with childhood SHS exposure only (aOR 1.34;
95% CI 0.83, 2.16) and with both childhood and adulthood SHS exposure (aOR 1.36; 95% CI
0.88, 2.11), but neither finding reached statistical significance. We found no evidence of a
significant difference in the association between lifetime composite smoking exposure by BC
subtype (Phet=0.15).
Among never smokers, neither childhood nor adulthood SHS exposure status was
significantly associated with BC risk overall of by subtype (Table 5.3). A significant difference
in the odds ratios for duration of childhood exposure and BC risk was observed by BC subtype
(Phet =0.01). No significant associations between childhood SHS exposure and BC risk were
observed by BC subtype, except a significantly decreased risk of HER2-type BC was observed
for ≥15 years of SHS exposure in childhood (aOR 0.31; 95% CI 0.11, 0.85). We observed an
increased, non-significant risk of TNBC with childhood SHS exposures, especially with ≥15
years of SHS exposure (aOR 1.42; 95% CI 0.92, 2.22). SHS exposure and duration of SHS
exposure in adulthood were not associated with BC risk, overall or by molecular BC subtype.
We evaluated the potential of effect modification; we did not identify a significant
interaction between prenatal smoke exposure status and the composite smoking or SHS variables
(Table 5.4). Race was also not a significant effect modifier of the associations between lifetime
composite smoking and SHS exposures with BC risk (Table 5.5). Although HHP was a
significant effect modifier of the association between childhood SHS exposure and BC risk, the
findings by category of HHP were not statistically significant (HHP ≥200%: aOR 0.79; 95% CI
0.59, 1.06; HHP <200%: aOR 1.31; 95% CI 0.80, 2.16; Pinteraction=0.07). Notably, the magnitude
of the OR estimates for each of the categories capturing childhood SHS exposure in the lifetime
composite exposure and duration of childhood SHS variables were higher in the stratum of

124
participants with an HHP <200% of the FPL compared to the estimates for the same categories
among participants with an HHP ≥200% of the FPL, although a statistically significant
association with BC risk was not observed in any of the strata.

5.5 Discussion
In this population of NHB and NHW young women we described a high prevalence of
SHS exposure (about 75%). We did not identify an association between lifetime composite
smoking status and BC risk, overall or by BC subtype. A significant difference in the association
between duration of childhood SHS exposure and BC risk differed by BC subtype among never
smokers in the study population. This is likely driven by the 69% significantly decreased risk of
HER2-type BC and the non-significant 42% increased risk of TNBC for non-smoking women
≥15 years of childhood SHS exposure. The sample of cases with HER2-type BC is small
compared to the other BC subtypes in the analysis among never smokers, therefore, this finding
must be interpreted with caution.
Ever smoking as defined in the lifetime composite smoking variable was not associated
with an increased risk of BC. Although the odds ratios did not reach statistical significance, the
magnitude of the odds ratios estimates for each of the BC subtypes are similar to the estimates in
our previous assessment of active smoking and BC risk (see Chapter 4). However, modeling ever
smoking in a lifetime composite exposure variable allowed for the estimation of the cumulative
effects of SHS and active smoke exposure using the same reference group and mutually
exclusive categories from our sample population. By doing so, we can evaluate which smoking
characteristic contributes to the greatest increase in BC risk. Although significant results were
not observed, these findings suggest that ever smoking contributes to an increased risk of BC

125
overall, of Luminal A BC, and of HER2-type BC. The findings suggest that childhood SHS is
associated with an increased risk of TNBC as the two composite smoke exposure categories that
included childhood SHS exposure both estimated about a 35% increased risk of TNBC.
Previous studies on the association between SHS exposure and BC risk have reported
conflicting results. Many studies have reported inconclusive findings (11-14; 19; 20). Yet other
studies have identified an increased BC risk with SHS exposure (10; 13; 21-23). For the
California Teachers study, statistically significant findings were only observed among
postmenopausal women and for moderate to high levels of SHS exposure, but was not observed
among premenopausal women (13). However, two subsequent large meta-analyses of SHS in
premenopausal populations found evidence of an association between SHS and BC risk. In 2014,
the Surgeon General reported that premenopausal women with secondhand smoke exposure were
at increased risk to develop BC than premenopausal women with no SHS exposure (relative risk
(RR) 1.14, 95% CI 1.06, 1.23) (24). Furthermore, Lee and Hamling provided additional support
for the Surgeon General’s finding with their own meta-analysis among nonsmoking women and
estimated a relative risk of 1.36 (95% CI 1.15, 1.60) among premenopausal women exposed to
passive smoking compared to those not exposed (25).
This study had many strengths. The analysis was conducted using data from a large,
racially diverse population-based case-control study that was designed to explore racial and
socioeconomic disparities in BC risk. We conducted a comprehensive structured survey to
capture multiple periods of personal smoking and SHS exposure to develop composite and SHS
exposure variables. Diagnoses of breast cancer were all histologically confirmed and identified
using the well-established local SEER registries to reduce any bias in assigning the outcome. An
extensive population-based sampling method was used to identify and recruit controls. Sample

126
weights to account for nonresponse and sampling bias were estimated and applied to all
statistical models. This study represents only the third study to explore the association between
smoking and BC subtype with a sample of HER2-type BCs among a population of young
women. To our knowledge, this is the only assessment of composite lifetime smoking exposures
and HER2-type BC risk among premenopausal women because routine reporting of HER2 status
in the SEER cancer registries only began in 2010 (26; 27).
Still there were some limitations in our study. We conducted the analysis using self-
reported smoking exposures which are subject to some differential recall bias. We attempted to
minimize this bias with the use of lifetime history calendars that were employed equally among
cases and controls to assist in recalling life events. Although we have information on the duration
of SHS exposures we do not have other details of exposures such as the source of the various
SHS exposures during childhood and adulthood. Thus, we were not able to assess workplace or
spousal SHS exposure. However, since multiple periods of SHS exposure could be reported we
anticipate that both types of exposure were captured.
In summary, our results suggest that personal smoking and childhood SHS exposures are
associated with an increased risk of BC. The risk varies by tumor subtype with a stronger
association between childhood SHS exposure and TNBC risk. Further research is warranted to
confirm these associations. However, these findings suggest that to reduce BC incidence in the
US, populations of NHB and NHW women may benefit from comprehensive smoking policies to
protect children from SHS exposure and programs to prevent smoking initiation or to encourage
early smoking cessation.

127
5.6 References
1. DeSantis, C. E., Fedewa, S. A., Goding Sauer, A., Kramer, J. L., Smith, R. A., & Jemal, A.
(2016). Breast cancer statistics, 2015: Convergence of incidence rates between black and
white women. CA Cancer J Clin, 66(1), 31-42. doi:10.3322/caac.21320
2. Gaudet, M. M., Carter, B. D., Brinton, L. A., Falk, R. T., Gram, I. T., Luo, J., . . . Gapstur,
S. M. (2017). Pooled analysis of active cigarette smoking and invasive breast cancer risk in
14 cohort studies. Int J Epidemiol, 46(3), 881-893. doi:10.1093/ije/dyw288
3. American Cancer Society. (2016). Cancer Facts & Figures for African Americans 2016-
2018. https://www.cancer.org/content/dam/cancer-org/research/cancer-facts-and-
statistics/cancer-facts-and-figures-for-african-americans/cancer-facts-and-figures-for-
african-americans-2016-2018.pdf
4. Liu, L., Deapen, D., & Bernstein, L. (1998). Socioeconomic status and cancers of the
female breast and reproductive organs: a comparison across racial/ethnic populations in
Los Angeles County, California (United States). Cancer Causes Control, 9(4), 369-380.
5. Palmer, J. R., Boggs, D. A., Wise, L. A., Adams-Campbell, L. L., & Rosenberg, L. (2012).
Individual and neighborhood socioeconomic status in relation to breast cancer incidence in
African-American women. Am J Epidemiol, 176(12), 1141-1146. doi:10.1093/aje/kws211
6. Vona-Davis, L., & Rose, D. P. (2009). The influence of socioeconomic disparities on
breast cancer tumor biology and prognosis: a review. J Womens Health (Larchmt), 18(6),
883-893. doi:10.1089/jwh.2008.1127
7. Centers for Disease Control and Prevention Office on Smoking and Health. (2019).
Smoking & Tobacco Use: African Americans and Tobacco Use.
https://www.cdc.gov/tobacco/disparities/african-americans/index.htm
8. Kandel, D., Schaffran, C., Hu, M. C., & Thomas, Y. (2011). Age-related differences in
cigarette smoking among whites and African-Americans: evidence for the crossover
hypothesis. Drug Alcohol Depend, 118(2-3), 280-287.
doi:10.1016/j.drugalcdep.2011.04.008
9. Tsai, J., Homa, D. M., Gentzke, A. S., Mahoney, M., Sharapova, S. R., Sosnoff, C. S., . . .
Trivers, K. F. (2018). Exposure to Secondhand Smoke Among Nonsmokers - United
States, 1988-2014. MMWR Morb Mortal Wkly Rep, 67(48), 1342-1346.
doi:10.15585/mmwr.mm6748a3
10. Dossus, L., Boutron-Ruault, M. C., Kaaks, R., Gram, I. T., Vilier, A., Fervers, B., . . .
Clavel-Chapelon, F. (2014). Active and passive cigarette smoking and breast cancer risk:
results from the EPIC cohort. Int J Cancer, 134(8), 1871-1888. doi:10.1002/ijc.28508
11. Xue, F., Willett, W. C., Rosner, B. A., Hankinson, S. E., & Michels, K. B. (2011).
Cigarette smoking and the incidence of breast cancer. Arch Intern Med, 171(2), 125-133.
doi:10.1001/archinternmed.2010.503

128
12. Roddam, A. W., Pirie, K., Pike, M. C., Chilvers, C., Crossley, B., Hermon, C., . . . Beral,
V. (2007). Active and passive smoking and the risk of breast cancer in women aged 36-45
years: a population based case-control study in the UK. Br J Cancer, 97(3), 434-439.
doi:10.1038/sj.bjc.6603859
13. Reynolds, P., Goldberg, D., Hurley, S., Nelson, D. O., Largent, J., Henderson, K. D., &
Bernstein, L. (2009). Passive smoking and risk of breast cancer in the California teachers
study. Cancer Epidemiol Biomarkers Prev, 18(12), 3389-3398. doi:10.1158/1055-
9965.EPI-09-0936
14. Reynolds, P., Hurley, S., Goldberg, D. E., Anton-Culver, H., Bernstein, L., Deapen, D., . . .
Ziogas, A. (2004). Active smoking, household passive smoking, and breast cancer:
evidence from the California Teachers Study. J Natl Cancer Inst, 96(1), 29-37.
15. Rosenberg, L., Boggs, D. A., Bethea, T. N., Wise, L. A., Adams-Campbell, L. L., &
Palmer, J. R. (2013). A prospective study of smoking and breast cancer risk among
African-American women. Cancer Causes Control, 24(12), 2207-2215.
doi:10.1007/s10552-013-0298-6
16. Provenzano, E., Ulaner, G. A., & Chin, S. F. (2018). Molecular Classification of Breast
Cancer. PET Clin, 13(3), 325-338. doi:10.1016/j.cpet.2018.02.004
17. United States Census Bureau. (2020, August 21, 2020). Poverty Thresholds. Retrieved
from https://www.census.gov/data/tables/time-series/demo/income-poverty/historical-
poverty-thresholds.html
18. Centers for Medicare & Medicaid Services. (n.d.). Basic Health Program. Retrieved from
https://www.medicaid.gov/basic-health-program/index.html
19. Rollison, D. E., Brownson, R. C., Hathcock, H. L., & Newschaffer, C. J. (2008). Case-
control study of tobacco smoke exposure and breast cancer risk in Delaware. BMC Cancer,
8, 157. doi:10.1186/1471-2407-8-157
20. Bonner, M. R., Nie, J., Han, D., Vena, J. E., Rogerson, P., Muti, P., . . . Freudenheim, J. L.
(2005). Secondhand smoke exposure in early life and the risk of breast cancer among never
smokers (United States). Cancer Causes Control, 16(6), 683-689. doi:10.1007/s10552-
005-1906-x
21. Luo, J., Margolis, K. L., Wactawski-Wende, J., Horn, K., Messina, C., Stefanick, M. L., . .
. Rohan, T. E. (2011). Association of active and passive smoking with risk of breast cancer
among postmenopausal women: a prospective cohort study. BMJ, 342, d1016.
doi:10.1136/bmj.d1016
22. Li, B., Wang, L., Lu, M. S., Mo, X. F., Lin, F. Y., Ho, S. C., & Zhang, C. X. (2015).
Passive Smoking and Breast Cancer Risk among Non-Smoking Women: A Case-Control
Study in China. PLoS One, 10(4), e0125894. doi:10.1371/journal.pone.0125894

129
23. Kropp, S., & Chang-Claude, J. (2002). Active and passive smoking and risk of breast
cancer by age 50 years among German women. Am J Epidemiol, 156(7), 616-626.
24. U.S. Department of Health and Human Services. (2014). Cancer. In U.S. Department of
Health and Human Services Centers for Disease Control and Prevention National Center
for Chronic Disease Prevention and Health Promotion Office on Smoking and Health
(Ed.), The Health Consequences of Smoking-50 Years of Progress: A Report of the
Surgeon General. Atlanta, GA.
25. Lee, P. N., & Hamling, J. S. (2016). Environmental tobacco smoke exposure and risk of
breast cancer in nonsmoking women. An updated review and meta-analysis. Inhal Toxicol,
28(10), 431-454. doi:10.1080/08958378.2016.1210701
26. Surveillance, E., and End Results (SEER) Program,. Breast Subtype (2010+). Retrieved
from https://seer.cancer.gov/seerstat/databases/ssf/breast-subtype.html
27. DeSantis, C. E., Ma, J., Gaudet, M. M., Newman, L. A., Miller, K. D., Goding Sauer, A., . .
. Siegel, R. L. (2019). Breast cancer statistics, 2019. CA Cancer J Clin, 69(6), 438-451.
doi:10.3322/caac.21583

130
Table 5.1. Sample weighted characteristics of the 1,381 controls in the Young Women’s Health History Study by prenatal and passive
smoke exposure status
a

Lifetime Composite Smoke Exposure
c

Total Population
b

No personal or
secondhand
exposure
Childhood
secondhand
exposure only
Adulthood
secondhand
exposure only
Both childhood
& adulthood
secondhand
exposures
Ever smoker p-value
N (W%) 1332 (100) 313 (28.7) 225 (16.1) 73 (5.7) 241 (16.0) 479 (33.5)

Region

0.001
Metropolitan Detroit 700 (60.7) 132 (23.8) 107 (13.7) 31 (5.4) 146 (18.1) 283 (38.9)

Los Angeles County 632 (39.4) 181 (36.2) 118 (19.7) 42 (6.2) 95 (12.6) 196 (25.3)

Age at diagnosis/reference year, years
weighted mean (95% CI)
34.3
(33.2-35.4)
30.5
(29.1-32.0)
35.3
(33.0-37.5)
32.3
(28.5-36.1)
36.9
(34.7-39.2)
36.1
(35.0-37.3)
<0.001
Age at diagnosis/reference year, years

<0.001
20-29 238 (36.6) 96 (44.2) 33 (15.0) 20 (6.8) 27 (9.3) 62 (24.7)

30-39 466 (30.1) 108 (22.2) 72 (14.8) 30 (6.5) 84 (18.7) 172 (37.9)

40-49 628 (33.3) 109 (17.7) 120 (18.4) 23 (3.8) 130 (20.8) 245 (39.3)

Race

0.09
Non-Hispanic White (NHW) 679 (64.5) 152 (28.9) 105 (14.1) 40 (5.8) 104 (14.5) 278 (36.7)

Non-Hispanic Black (NHB) 653 (35.5) 161 (28.3) 120 (19.6) 33 (5.5) 137 (18.6) 201 (27.8)

Household poverty level (HHP)

0.08
≥200% federal poverty level 691 (53.4) 182 (32.0) 145 (18.2) 41 (5.3) 111 (16.7) 212 (27.8)

<200% federal poverty level 601 (43.7) 121 (24.3) 76 (14.2) 28 (5.3) 126 (15.4) 249 (40.7)

Missing 40 (2.9) 10 (33.6) 4 (5.3) 4 (19.0) 4 (10.9) 18 (31.1)

Education

<0.001
High school diploma or less 286 (18.3) 51 (19.7) 35 (12.8) 12 (5.2) 64 (19.2) 124 (43.1)

Vocational school, associate's
degree, or some college
541 (43.4) 99 (23.0) 83 (15.2) 22 (4.2) 112 (19.1) 224 (38.4)

Bachelor's degree or higher 505 (38.3) 163 (39.5) 107 (18.6) 39 (7.7) 65 (10.8) 131 (23.4)

Adult Body Mass Index, kg/m
2

0.23
Underweight 38 (4.7) 12 (49.5) 4 (5.4) 2 (3.5) 4 (11.4) 16 (30.3)

Normal 468 (38.1) 134 (33.3) 83 (17.9) 27 (6.8) 52 (10.9) 172 (31.0)

Overweight 371 (28.5) 88 (26.4) 67 (18.0) 18 (4.8) 69 (15.1) 129 (35.8)

Obese 455 (28.7) 79 (21.5) 71 (13.4) 26 (5.5) 116 (24.2) 162 (35.2)

Age at menarche, years (mean ± SD)

0.23
≤11 392 (31.4) 84 (22.6) 74 (22.6) 18 (6.0) 82 (17.3) 134 (31.5)

12 405 (31.8) 95 (33.8) 66 (10.2) 21 (5.1) 61 (14.3) 161 (36.6)

13 291 (20.6) 76 (29.3) 38 (14.7) 19 (5.5) 51 (14.8) 107 (35.7)

≥14 244 (16.2) 58 (29.8) 47 (16.7) 15 (6.8) 47 (18.1) 77 (28.7)

Menopausal Status

0.006

131
Lifetime Composite Smoke Exposure
c

Total Population
b

No personal or
secondhand
exposure
Childhood
secondhand
exposure only
Adulthood
secondhand
exposure only
Both childhood
& adulthood
secondhand
exposures
Ever smoker p-value
Premenopausal 1,181 (92.0) 295 (30.0) 199 (16.1) 65 (5.8) 214 (15.8) 407 (32.3)

Peri-/postmenopausal 151 (8.0) 18 (13.8) 26 (15.6) 8 (5.0) 27 (18.3) 72 (47.4)

Joint parity & age at first birth
ⱡ
, years

0.007
Nulliparous 380 (38.2) 117 (38.2) 49 (14.9) 28 (7.0) 61 (13.8) 125 (26.3)

1-2, <25 294 (20.0) 55 (25.3) 53 (15.0) 13 (4.8) 55 (15.9) 117 (38.9)

1-2, ≥25 309 (21.4) 73 (24.6) 70 (20.3) 18 (5.5) 52 (15.5) 96 (34.2)

3+, <25 267 (15.4) 43 (15.1) 35 (12.9) 10 (4.3) 58 (20.4) 121 (47.4)

3+, ≥25 82 (5.0) 25 (30.0) 18 (21.5) 4 (4.8) 15 (21.6) 20 (22.1)

Family history of breast cancer

0.33
No 1,155 (88.2) 276 (30.0) 204 (16.4) 66 (5.9) 201 (15.3) 407 (32.3)

Yes 111 (7.4) 24 (20.0) 15 (14.4) 4 (4.4) 26 (19.0) 42 (42.2)

Missing 66 (4.4) 13 (17.0) 6 (11.5) 3 (3.2) 14 (24.2) 30 (44.1)

Cumulative average lifetime alcohol
consumption, g/day

<0.001
0 (Abstainers) 425 (30.3) 132 (34.7) 81 (18.8) 24 (7.3) 102 (20.6) 85 (18.5)

0.1 - 6.9 269 (68.8) 93 (27.9) 91 (19.7) 21 (4.1) 64 (17.1) 124 (31.2)

7 - 13.9 212 (15.7) 48 (29.3) 28 (15.4) 13 (6.3) 35 (13.7) 88 (35.3)

14 - 27.9 184 (14.8) 33 (28.2) 21 (12.1) 11 (5.9) 26 (11.3) 93 (42.4)

≥ 28 118 (8.9) 7 (11.1) 4 (2.1) 4 (4.4) 14 (7.8) 89 (74.6)

a
n=49 participants missing smoking status not included. All values are absolute frequencies and sample weighted row percentages unless otherwise specified. Missing categories
are not included in chi-square p-value estimates.
b
Absolute frequencies and sample weighted column percentages are presented for the total population.
c
n=1 participant missing lifetime composite smoke exposure status not included
Abbreviations: CI, confidence interval; g, grams; kg, kilograms; m, meter.

132
Table 5.2. Lifetime smoking characteristics of the 3,193 women in the Young Women’s Health History Study by breast cancer status
and subtype
a

Breast Cancer Status Breast Cancer Subtype
b

Control Case Luminal A Luminal B HER2-type TNBC
N (W%) 1,332 (50.1%) 1,725 (49.9%) 666 (21.3%) 538 (15.2%) 102 (2.7%) 294 (7.4%)
Prenatal smoke exposure
No 829 (59.8) 1,087 (62.1) 432 (65.1) 354 (62.4) 62 (60.8) 171 (58.8)
Yes 238 (24.7) 394 (25.1) 147 (23.3) 119 (26.5) 24 (23.8) 65 (22.9)
Unknown 264 (15.3) 241 (12.6) 85 (11.4) 64 (11.1) 16 (15.4) 58 (18.3)
Missing 1 (0.2) 3 (0.1) 2 (0.2) 1 (0.1) 0 (0.0) 0 (0.0)
Lifetime SHS exposure

No 354 (22.3) 434 (23.8) 166 (24.4) 164 (27.1) 21 (19.9) 56 (18.6)
Yes 977 (77.7) 1,290 (76.2) 499 (75.5) 374 (72.9) 81 (80.1) 238 (81.4)
Missing 1 (0.1) 1 (0.05) 1 (0.1) 0 (0.0) 0 (0.0) 0 (0.0)
Duration of lifetime SHS exposure
No 354 (22.3) 434 (23.8) 166 (24.4) 164 (27.1) 21 (19.9) 56 (18.6)
<15 years SHS exposure 385 (25.0) 459 (26.2) 175 (26.1) 146 (26.7) 27 (24.8) 80 (27.4)
≥15 years SHS exposure 588 (52.5) 827 (49.8) 321 (49.2) 227 (46.0) 54 (55.4) 158 (53.9)
Missing 5 (0.3) 5 (0.2) 4 (0.4) 1 (0.2) 0 (0.0) 0 (0.0)
Childhood SHS exposure
No 497 (32.3) 603 (34.0) 231 (35.1) 211 (36.1) 34 (32.1) 88 (30.1)
Yes 834 (67.6) 1,118 (65.8) 432 (64.5) 326 (63.8) 68 (67.9) 206 (69.9)
Missing 1 (0.1) 4 (0.2) 3 (0.4) 1 (0.1) 0 (0.0) 0 (0.0)
Duration of childhood SHS exposure
No 497 (32.3) 603 (34.0) 231 (35.1) 211 (36.1) 34 (32.1) 88 (30.1)
<15 years SHS exposure 426 (30.3) 528 (29.6) 199 (29.1) 167 (30.4) 36 (33.4) 85 (28.3)
≥15 years SHS exposure 401 (36.9) 580 (35.8) 227 (34.8) 158 (33.2) 30 (32.1) 121 (41.7)
Missing 8 (0.5) 14 (0.7) 9 (1.1) 2 (0.3) 2 (2.4) 0 (0.0)
Adulthood SHS exposure
No 641 (44.1) 834 (46.8) 331 (47.4) 289 (52.8) 36 (33.4) 132 (43.6)
Yes 691 (55.9) 891 (53.2) 335 (52.6) 249 (47.2) 66 (66.7) 162 (56.4)
Duration of adulthood SHS exposure
No 641 (44.1) 834 (46.8) 331 (47.4) 289 (52.8) 36 (33.4) 132 (43.6)
<10 years SHS exposure 368 (26.4) 416 (25.1) 154 (24.5) 129 (24.2) 28 (27.5) 64 (22.5)
≥10 years SHS exposure 322 (29.5) 473 (28.0) 181 (28.1) 119 (22.8) 38 (39.1) 97 (33.7)
Missing 1 (0.04) 2 (0.1) 0 (0.0) 1 (0.1) 0 (0.0) 1 (0.3)
Lifetime composite smoke exposure
No personal or secondhand exposure 313 (19.4) 369 (20.3) 143 (20.9) 136 (22.8) 19 (18.0) 51 (16.8)
Secondhand smoke exposure only

133
Breast Cancer Status Breast Cancer Subtype
b

Control Case Luminal A Luminal B HER2-type TNBC
In childhood only 225 (17.7) 306 (16.9) 118 (15.6) 98 (19.0) 13 (11.6) 60 (19.6)
In adulthood only 73 (4.8) 83 (5.2) 34 (5.8) 27 (5.4) 4 (3.3) 14 (4.5)
In childhood & adulthood 241 (19.9) 315 (18.1) 107 (16.4) 91 (16.8) 14 (12.5) 68 (23.1)
Ever smoker 479 (38.2) 649 (39.4) 262 (41.0) 185 (36.0) 52 (54.7) 101 (36.0)
Missing 1 (0.1) 3 (0.1) 2 (0.3) 1 (0.1) 0 (0.0) 0 (0.0)
a
All values are absolute frequencies and sample weighted column percentages. Data for n=136 participants missing smoking status not included.
b
Data for n=125 participants missing breast cancer subtype not included in analyses by breast cancer subtype.
Abbreviations: HER2, human epidermal growth factor receptor 2; TNBC, triple negative breast cancer.

134
Table 5.3. Association lifetime smoking exposures and breast cancer risk, overall and by tumor subtype
Overall Luminal A Luminal B HER2-type TNBC Pheterogeneity

Multivariable-
adjusted OR (95%
CI)
a

Multivariable-
adjusted OR (95%
CI)
a

Multivariable-
adjusted OR (95%
CI)
a

Multivariable-
adjusted OR (95%
CI)
a

Multivariable-
adjusted OR (95%
CI)
a

Total Population

Lifetime composite smoke
exposure

0.15
No personal or SHS exposure 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

Secondhand smoke exposure
only

In childhood only 0.93 (0.70-1.23) 0.80 (0.55-1.15) 1.00 (0.68-1.46) 0.70 (0.31-1.59) 1.34 (0.83-2.16)

In adulthood only 1.02 (0.64-1.61) 1.10 (0.62-1.96) 0.97 (0.54-1.74) 0.74 (0.20-2.72) 1.09 (0.54-2.17)

In childhood & adulthood 0.92 (0.68-1.25) 0.78 (0.53-1.16) 0.84 (0.56-1.27) 0.62 (0.29-1.31) 1.36 (0.88-2.11)

Ever smoker 1.14 (0.86-1.51) 1.15 (0.83-1.59) 1.00 (0.67-1.48) 1.46 (0.81-2.62) 1.15 (0.77-1.71)

Among Never Smokers –
Model 1
b

Childhood SHS smoke exposure

0.07
No 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

Yes 0.93 (0.72-1.18) 0.78 (0.59-1.03) 0.96 (0.68-1.35) 0.76 (0.37-1.56) 1.28 (0.87-1.87)

Adulthood SHS smoke exposure

No 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 0.79
Yes 0.98 (0.77-1.25) 1.01 (0.73-1.39) 0.85 (0.59-1.24) 0.77 (0.38-1.56) 1.00 (0.68-1.47)

Among Never Smokers –
Model 2
b

Duration of childhood SHS
exposure

0.01
No 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

<15 years SHS exposure 0.96 (0.72-1.28) 0.76 (0.53-1.09) 1.03 (0.69-1.54) 1.24 (0.60-2.56) 1.11 (0.69-1.76)

≥15 years SHS exposure 0.87 (0.66-1.15) 0.75 (0.54-1.04) 0.89 (0.60-1.31) 0.31 (0.11-0.85) 1.42 (0.92-2.22)

Duration of adulthood SHS
exposure

0.93
No 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

<10 years SHS exposure 0.94 (0.71-1.25) 0.94 (0.66-1.35) 0.85 (0.58-1.26) 0.87 (0.38-2.02) 0.83 (0.53-1.31)

≥10 years SHS exposure 1.13 (0.77-1.66) 1.23 (0.73-2.06) 0.89 (0.50-1.58) 0.97 (0.38-2.49) 1.25 (0.72-2.18)

a
Adjusted for study site (Detroit, Los Angeles); age at diagnosis; household poverty (≥200%, <200% of FPL); first degree family member w/ BC (Y, N, unknown); age at
menarche (≤11, 12, 13, ≥14 years); BMI 12 months before reference date (underweight, normal, overweight, obese); alcohol use (0, 0.1-6.9, 7-13.9, 14-42 and >42 grams/day); age
at first birth/parity (nulliparous, 1-2 children and <25 years, 1-2 children and 25-29 years, 1-2 children 30+ years, 3+ children and <25 years, 3+ children and 25-29 years, and 3+
children and 30+ years); premenopausal status (Y, N).
b
Mutually adjusted for each smoking characteristic.

135
Abbreviations: BC, breast cancer; BMI, body mass index; CI, confidence interval; ref, reference; FPL, federal poverty level; FTP, full-term pregnancy; HER2, human epidermal
growth factor receptor 2; N, no; OR, odds ratio; TNBC, triple negative breast cancer; Y, yes.
Bold values significant at p<0.05

136
Table 5.4. Association between lifetime smoking history and breast cancer risk among women without prenatal smoke exposure
No prenatal smoke exposure With prenatal smoke exposure

Crude OR (95% CI)
Multivariable-adjusted
OR (95% CI)
a

Crude OR (95% CI)
Multivariable-adjusted
OR (95% CI)
a

Pinteraction
Total Population

Lifetime composite smoke exposure

0.80
No personal or secondhand exposure 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

Secondhand smoke exposure only
In childhood only 0.92 (0.64-1.32) 0.95 (0.65-1.39) 0.61 (0.16-2.30) 0.42 (0.11-1.56)

In adulthood only 0.98 (0.59-1.64) 0.98 (0.58-1.65) 0.36 (0.01-9.59) 0.14 (0.01-3.56)

In childhood & adulthood 0.88 (0.63-1.22) 0.94 (0.65-1.36) 0.70 (0.18-2.65) 0.50 (0.14-1.79)

Ever smoker 0.92 (0.66-1.28) 1.10 (0.76-1.60) 0.81 (0.23-2.82) 0.65 (0.19-2.19)

Among Never Smokers - Model 1
b

Childhood SHS exposure

0.98
No 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

Yes 0.91 (0.69-1.20) 0.94 (0.69-1.27) 0.77 (0.23-2.59) 0.92 (0.19-4.32)

Adulthood SHS exposure

0.58
No 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

Yes 0.96 (0.72-1.30) 0.98 (0.71-1.36) 1.10 (0.62-1.95) 0.95 (0.50-1.78)

Among Never Smokers - Model 2
b

Duration of childhood SHS exposure

0.57
No 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

<15 years SHS exposure 0.89 (0.63-1.26) 0.93 (0.64-1.34) 1.11 (0.30-4.06) 1.26 (0.25-6.41)

≥15 years SHS exposure 0.88 (0.60-1.30) 0.87 (0.58-1.32) 0.69 (0.21-2.29) 0.80 (0.17-3.78)

Duration of adulthood SHS exposure

0.85
No 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)

<10 years SHS exposure 0.91 (0.63-1.30) 0.89 (0.61-1.31) 1.04 (0.56-1.91) 0.87 (0.44-1.74)

≥10 years SHS exposure 1.14 (0.75-1.74) 1.31 (0.83-2.07) 1.36 (0.62-2.97) 1.24 (0.52-2.96)

a
Adjusted for study site (Detroit, Los Angeles); age at diagnosis; household poverty (≥200%, <200% of FPL); first degree family member w/ BC (Y, N, unknown); age at
menarche (≤11, 12, 13, ≥14 years); BMI 12 months before reference date (underweight, normal, overweight, obese); alcohol use (0, 0.1-6.9, 7-13.9, 14-42 and >42 grams/day); age
at first birth/parity (nulliparous, 1-2 children and <25 years, 1-2 children and 25-29 years, 1-2 children 30+ years, 3+ children and <25 years, 3+ children and 25-29 years, and 3+
children and 30+ years); premenopausal status (Y, N). Models stratified by SEP are not adjusted for SEP.
b
Mutually adjusted for each smoking characteristic.
Abbreviations: BC, breast cancer; BMI, body mass index; CI, confidence interval; FPL, federal poverty level; FTP, full-term pregnancy; HHP, household poverty; N, no; OR, odds
ratio; ref, reference; SEP, socioeconomic position; Y, yes.

137
Table 5.5. Association between lifetime smoking history and breast cancer risk by race and SEP
Non-Hispanic White Non-Hispanic Black

HHP ≥200% HHP <200%

Multivariable-adjusted
OR (95% CI)
a

Multivariable-adjusted
OR (95% CI)
a

Pinteraction
Multivariable-adjusted
OR (95% CI)
a

Multivariable-adjusted
OR (95% CI)
a

Pinteraction
Total Population

Lifetime composite smoke exposure

0.62

0.32
No personal or secondhand exposure 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

Secondhand smoke exposure only
In childhood only 0.82 (0.54-1.24) 0.98 (0.66-1.45)

0.77 (0.53-1.12) 1.47 (0.87-2.48)

In adulthood only 0.89 (0.50-1.60) 1.30 (0.66-2.54)

0.86 (0.46-1.59) 1.63 (0.65-4.07)

In childhood & adulthood 0.92 (0.60-1.40) 0.91 (0.63-1.32)

0.74 (0.49-1.11) 1.29 (0.84-1.97)

Ever smoker 1.23 (0.83-1.83) 0.98 (0.67-1.42)

1.06 (0.70-1.60) 1.38 (0.93-2.06)

Among Never Smokers - Model 1
b

Childhood SHS exposure

0.80

0.07
No 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

Yes 0.90 (0.62-1.31) 0.94 (0.67-1.34)

0.79 (0.59-1.06) 1.31 (0.80-2.16)

Adulthood SHS exposure

0.92

0.40
No 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

Yes 1.03 (0.74-1.45) 0.92 (0.64-1.33)

0.91 (0.64-1.29) 0.94 (0.59-1.50)

Among Never Smokers - Model 2
b

Duration of childhood SHS exposure

0.43

0.22
No 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<15 years SHS exposure 0.97 (0.62-1.53) 0.88 (0.60-1.29)

0.82 (0.57-1.18) 1.26 (0.74-2.14)

≥15 years SHS exposure 0.83 (0.55-1.26) 1.00 (0.67-1.48)

0.75 (0.54-1.05) 1.34 (0.72-2.47)

Duration of adulthood SHS exposure

0.52

0.28
No 1.00 (ref) 1.00 (ref)

1.00 (ref) 1.00 (ref)

<10 years SHS exposure 1.05 (0.72-1.53) 0.75 (0.46-1.21)

0.95 (0.65-1.39) 0.74 (0.44-1.26)

≥10 years SHS exposure 1.08 (0.57-2.04) 1.19 (0.73-1.93)

0.87 (0.48-1.57) 1.22 (0.70-2.12)

a
Adjusted for study site (Detroit, Los Angeles); age at diagnosis; household poverty (≥200%, <200% of FPL); first degree family member w/ BC (Y, N, unknown); BMI 12
months before reference date (underweight, normal, overweight, obese); alcohol use (0, 0.1-6.9, 7-13.9, 14-42 and >42 grams/day); age at first birth/parity (nulliparous, 1-2
children and <25 years, 1-2 children and 25-29 years, 1-2 children 30+ years, 3+ children and <25 years, 3+ children and 25-29 years, and 3+ children and 30+ years);
premenopausal status (Y, N). Models stratified by SEP are not adjusted for SEP.
b
Mutually adjusted for each smoking characteristic.

138
Abbreviations: BC, breast cancer; BMI, body mass index; CI, confidence interval; FPL, federal poverty level; FTP, full-term pregnancy; HHP, household poverty; N, no; OR, odds
ratio; ref, reference; SEP, socioeconomic position; Y, yes.

139
CHAPTER SIX: Summary and Future Research

6.1 Summary and Conclusion
This dissertation was designed to examine the association between smoking and hormone
levels among young women and to assess the association between multiple smoking
characteristics and breast cancer (BC) risk among a population of young, non-Hispanic Black
and White women.
In Chapter 2, a systematic review and meta-analysis was developed to describe the
association between smoking and mean endogenous hormone levels in a population of
premenopausal women. Smoking status was associated with high mean androgen levels but was
not associated with differences in standardized mean estradiol levels in this meta-analysis of
premenopausal women. Moderate and substantial heterogeneity was observed for the summary
estimates within menstrual phase for most analytes. However, heterogeneity was not explained
by the year of publication, whether the analyte measures were covariate adjusted or not or the
region of origin. Also, for estradiol, having measures by urine or blood biospecimen types did
not explain the heterogeneity in the sample.
Chapter 3 introduced the design of the case-control study that is used for the remainder of
the analysis, the Young Women’s Health History Study. In Chapter 4, the details of the analysis
of personal smoking history and breast cancer risk are explored. In summary, an association
between ever smoking and breast cancer risk was not observed overall, however multiple
personal smoking factors were positively associated with Luminal A and Human Epidermal
Growth Factor Receptor 2 (HER2)-type breast cancer, including ever smoking, pack-years of
smoking, the average number of cigarettes smoked per day and the years since smoking

140
initiation. Smoking before first full-term pregnancy (FFTP) was associated with Luminal A BC,
and age smoking initiated was significantly associated with an increase in HER2-type BC risk.
Chapter 5 includes a continuation of our exploration of the association between smoking
and breast cancer risk with a description of secondhand smoke (SHS) exposure and lifetime
composite smoking exposure in the population. We described a high prevalence of secondhand
smoke exposures in the population (~75% of participants with SHS exposure in childhood or
adulthood). A 69% significantly decreased risk of HER2-type BC was observed for ≥15 years of
SHS exposure in childhood and a non-significant, increased risk of triple negative BC (TNBC)
was suggested with childhood SHS exposure in multiple models. We did not identify a
significant association between lifetime composite smoking exposures and breast cancer risk,
overall or by breast cancer subtype.

6.2 Implications and Future Directions
Collectively this dissertation explored some gaps in the existing literature. In our
assessments by menstrual phase, we observed a decrease in mean estrogen levels in the luteal
phase; smoking was associated with an increased mean estrogen level in the follicular phase
compared to non-smokers. This was an unexpected finding that warrants further evaluation in
larger populations with carefully timed hormone measures. Furthermore, this analysis
highlighted the need for covariate adjusted results when assessing hormone levels among
premenopausal women. Body mass index and physical activity levels are also associated with
differences in hormone levels and could be associated with smoking, however, these covariates
were not adjusted for in many of the studies included in the meta-analysis (1). Overall, these
findings may suggest that smoking does not contribute to premenopausal breast carcinogenesis

141
via disruption of endogenous hormone levels and receptor binding, but since we did not evaluate
the hormone metabolites, we cannot make inferences on how smoking may influence the full
metabolic pathway (2).
In the analysis of personal smoking history, our results indicate that ever smoking,
heavier, and long-term smoking before FFTP, were associated with an increased risk of Luminal
A and HER2-type BC in young women. Previous studies have evaluated the association between
smoking and BC risk by subtype, and many have identified a positive association between
smoking and risk of Luminal A or hormone receptor positive BC, however, only two previous
studies have been conducted among younger or premenopausal women. Our findings are
consistent with the increased risk of BC found among premenopausal smokers in the Morabia
and Kawai studies (3; 4).
Although, other studies have identified an association between smoking and Luminal A
breast cancer and triple negative breast cancer, this may be the first study to identify an
association between smoking and HER2-type breast cancer risk. The association between
smoking and HER2-type breast cancer was consistently observed in this analysis, however
additional studies are needed in populations with larger HER2-type samples to confirm the
association.
In terms of the mechanisms of carcinogenesis, altered metabolic activation may
contribute to breast carcinogenesis (2). In 2003, the Long Island Breast Cancer Study identified a
significant positive association between smoking and DNA adducts in normal breast tissues (5).
In the case of HER2-type breast cancer, if smoking carcinogenesis is following a pathway of
altered metabolic activation, the formation of the DNA adducts may lead to DNA miscoding that
contributes to the over-expression of the HER2 protein and impaired tumor suppression (2). A

142
European study of baseline concentrations of aromatic DNA adducts in the population found that
there was as 60% increased risk of BC with increasing DNA adduct concentration in white blood
cells (RR 1.61; 95% CI 1.29-2.01) (6). They also detected a significant interaction with a higher
effect of DNA adducts on breast cancer risk among former (RR 2.89; 95% CI 1.42-5.86) and
current smokers (RR 2.19; 95% CI 1.22-3.93), therefore adding further support for the
plausibility of this pathway (6). The European study also considered BC subtype but only by ER
status. Additional studies of sex-hormone metabolism and DNA adducts in breast tissue are
needed in populations of smokers compared to non-smokers to understand more about smoking
carcinogenesis in Luminal A, HER2-type and TN BCs.
Based on the findings of this research, and consistent with other studies, education in
reducing secondhand smoking exposures, effective smoking cessation programs, and efforts in
preventing smoking initiation early in life are necessary to reduce BC incidence among young
women in the US.

143
6.3 References
1. Endogenous Hormones Breast Cancer Collaborative Group, Key, T. J., Appleby, P. N.,
Reeves, G. K., Roddam, A. W., Helzlsouer, K. J., . . . Strickler, H. D. (2011). Circulating
sex hormones and breast cancer risk factors in postmenopausal women: reanalysis of 13
studies. Br J Cancer, 105(5), 709-722. doi:10.1038/bjc.2011.254
2. U.S. Department of Health and Human Services. (2010). How Tobacco Smoke Causes
Disease: The Biology and Behavioral Basis for Smoking-Attributable Disease: A Report of
the Surgeon General. Atlanta (GA): U.S. Department of Health and Human Services,
Centers for Disease Control and Prevention, National Center for Chronic Disease
Prevention and Health Promotion, Office on Smoking and Health.
3. Kawai, M., Malone, K. E., Tang, M. T., & Li, C. I. (2014). Active smoking and the risk of
estrogen receptor-positive and triple-negative breast cancer among women ages 20 to 44
years. Cancer, 120(7), 1026-1034. doi:10.1002/cncr.28402
4. Morabia, A., Bernstein, M., Ruiz, J., Heritier, S., Diebold Berger, S., & Borisch, B. (1998).
Relation of smoking to breast cancer by estrogen receptor status. Int J Cancer, 75(3), 339-
342.
5. Faraglia, B., Chen, S. Y., Gammon, M. D., Zhang, Y., Teitelbaum, S. L., Neugut, A. I., . . .
Santella, R. M. (2003). Evaluation of 4-aminobiphenyl-DNA adducts in human breast
cancer: the influence of tobacco smoke. Carcinogenesis, 24(4), 719-725.
doi:10.1093/carcin/bgg013
6. Agudo, A., Peluso, M., Munnia, A., Lujan-Barroso, L., Barricarte, A., Amiano, P., . . .
Gonzalez, C. A. (2017). Aromatic DNA adducts and breast cancer risk: a case-cohort study
within the EPIC-Spain. Carcinogenesis, 38(7), 691-698. doi:10.1093/carcin/bgx047

144

APPENDIX

145

Appendix 1. YWHHS questionnaire items used to derive tobacco exposure variables

A1.1 Personal smoking questions
• Have you smoked a total of 100 cigarettes in your lifetime?
• Before [REFERENCE DATE], did you ever smoke at least one cigarette a day for 6
months or longer?
• How old were you when you first started smoking cigarettes on a regular basis?
• At what age did you last stop smoking cigarettes?
• Thinking about the years between [REFERENCE AGE] and IF STILL SMOKING:
[REFERENCE DATE/ AGE], was there ever a period of 1 year or more in which you did
not smoke cigarettes?
• For how many years between [REFERENCE AGE] and IF STILL SMOKING:
[REFERENCE DATE/ AGE], did you not smoke cigarettes?
• (During the periods when you smoked,) how many cigarettes (do / did) you usually
smoke per day, per week, or per month? One package contains 20 cigarettes.

A1.2 Prenatal tobacco exposure questions
• From the study questionnaire completed by the participant: When she was pregnant with
you, did your mother smoke cigarettes?
• From the caregiver survey completed by the participant’s primary childhood caregiver:
Did you smoke during your pregnancy with your daughter?

146
A1.3 Secondhand smoke exposure questions
• Before you were 18 years old, were there any periods of time when your parent or
caregiver who took care of you most of the time smoked regularly in the house, or in your
presence? By regularly, we mean at least one cigarette per day, for six months or longer.
• How old were you when your parent or caregiver first started smoking regularly (at least
one cigarette per day for six months or longer) in the house or in your presence?
• How old were you when your daily exposure to this cigarette smoke first stopped for a
year or longer? (This could be because your caregiver stopped smoking, moved out of the
house or because you moved out.)
• For this question we are only asking about your exposure to cigarette smoke before you
were 18 years old. I will ask about your exposure from age 18 to [REFERENCE DATE],
later. Keeping this in mind, would you say you were exposed to cigarette smoke before
you were 18 years old?
• From the time you were born up to age 18, were there any periods of time lasting six
months or longer, when someone other than your parent or caregiver smoked at least one
cigarette per day, in your presence?
• How old were you when someone other than your parent or caregiver first started
smoking regularly (at least one cigarette per day for six months or longer) in the house or
in your presence?
• How old were you when your daily exposure to others' cigarette smoke first stopped for a
year or longer?
• For this question, we are only asking about your exposure to cigarette smoke before you
were 18 years old. I will ask about your exposure from age 18 to [REFERENCE DATE],

147
later. Keeping this in mind, would you say you were exposed to cigarette smoke before
you were 18 years old?
• From the age of 18 up to 12 months prior to [REFERENCE DATE], were there any
periods of time when lasting six months or longer when at least one cigarette per day was
smoked in your presence?
• How old were you when others first started smoking regularly (at least one cigarette per
day for six months or longer) in the house or in your presence?
• How old were you when your daily exposure to others’ cigarette smoke next stopped for
a year or longer?
• From the age of [REFERENCE AGE] up to 12 months prior to [REFERENCE DATE],
were there any other periods of time when lasting six months or longer when at least one
cigarette per day was smoked in your presence?

Abstract (if available)

Abstract Breast cancer is the most common form of cancer diagnosed among women ages 20 to 49 years in the United States. While multiple risk factors for breast cancer have been identified, the evidence of the role of tobacco exposure in breast cancer etiology has been conflicting. Recent studies among premenopausal women have suggested that smoking is associated with an increased risk of breast cancer. Still, additional studies are required to confirm this association among populations of young women (under age 50 years), with a description of findings by breast cancer subtype, and to evaluate differences in the association by race and socioeconomic position (SEP). This dissertation aims to add to the growing literature evaluating the association between tobacco exposure and breast cancer risk among young women, and on the potential effect that tobacco exposure may have on endogenous hormone levels.
Background information on breast cancer incidence, mortality and risk factors are provided in Chapter 1. Notable differences in breast cancer epidemiology are evident when stratified by age (i.e., women age under vs over 50 years), by race, and by SEP. These observations suggest differences in risk factors between women of a younger versus older age at breast cancer diagnosis and provide context for the results of the analyses conducted in this dissertation.
Three papers are included in this document: 1) a systematic review and meta-analysis of endogenous hormone levels by smoking status among premenopausal women

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The effects of hormonal exposures on ovarian and breast cancer risk

PDF

Genes and environment in prostate cancer risk and prognosis

PDF

Genes and hormonal factors involved in the development or recurrence of breast cancer

PDF

Identifying genetic, environmental, and lifestyle determinants of ethnic variation in risk of pancreatic cancer

PDF

Predictive factors of breast cancer survival: a population-based study

PDF

Screening and association testing of coding variation in steroid hormone coactivator and corepressor genes in relationship with breast cancer risk in multiple populations

PDF

The role of heritability and genetic variation in cancer and cancer survival

PDF

Arm lymphedema in a multi-ethnic cohort of female breast cancer survivors

PDF

Genetic and environmental risk factors for childhood cancer

PDF

Carcinogenic exposures in racial/ethnic groups

PDF

Associations between isoflavone soy protein (ISP) supplementation and breast cancer in postmenopausal women in the Women’s Isoflavone Soy Health (WISH) clinical trial

PDF

Multilevel sociodemographic correlates of the health and healthcare utilization of childhood cancer survivors

PDF

Pathogenic variants in cancer predisposition genes and risk of non-breast multiple primary cancers in breast cancer patients

PDF

A randomized, double-blind, placebo-controlled phase II trial to examine the effect of Polyphenon E on endogenous hormone levels

PDF

The role of depression symptoms on social information processing and tobacco use among adolescents

PDF

The influence of DNA repair genes and prenatal tobacco exposure on childhood acute lymphoblastic leukemia risk: a gene-environment interaction study

PDF

Effects of mint, menthol, and tobacco-flavored e-cigarettes on appeal and sensory effects, tobacco withdrawal…

PDF

An analysis of disease-free survival and overall survival in inflammatory breast cancer

PDF

Survival trends and related outcomes of survivors of childhood and young adult cancer

PDF

Instability of heart rate and rating of perceived exertion during high-intensity interval training in breast cancer patients undergoing anthracycline chemotherapy

Asset Metadata

Creator Ihenacho, Ugonna Nnebuaku (author)

Core Title The effects of tobacco exposure on hormone levels and breast cancer risk among young women

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Epidemiology

Publication Date 03/22/2021

Defense Date 02/26/2021

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

Tag active smoking,androgen,breast cancer,dehydroepiandrosterone,dehydroepiandrosterone sulfate,endogenous hormones,Epidemiology,Estrogen,human epidermal growth factor receptor 2-type breast cancer,lifetime smoking exposure,Luminal A breast cancer,Luminal B breast cancer,meta-analysis,molecular subtype,OAI-PMH Harvest,passive smoking,premenopausal,prenatal smoke exposure,progesterone,racial disparities,secondhand smoke exposure,sex hormone-binding globulin,sex hormones,Smoking,socioeconomic disparities,systematic review,testosterone,tobacco exposure,triple negative breast cancer,Young women

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Hamilton, Ann (committee chair), Wu, Anna (committee chair), Mack, Wendy (committee member), Press, Michael (committee member), Unger, Jennifer (committee member)

Creator Email ihenacho@usc.edu,unihenacho@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-429835

Unique identifier UC11667456

Identifier etd-IhenachoUg-9341.pdf (filename),usctheses-c89-429835 (legacy record id)

Legacy Identifier etd-IhenachoUg-9341.pdf

Dmrecord 429835

Document Type Dissertation

Rights Ihenacho, Ugonna Nnebuaku

Type texts

Source 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 a...

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