Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Mar 2.
Published in final edited form as: Environ Int. 2022 Jan 18;161:107088. doi: 10.1016/j.envint.2022.107088

Outdoor ambient air pollution and breast cancer survival among California participants of the Multiethnic Cohort Study

Iona Cheng a,b,*, Juan Yang a, Chiuchen Tseng c, Jun Wu d, Shannon M Conroy e, Salma Shariff-Marco a,b, Scarlett Lin Gomez a,b, Alice S Whittemore f, Daniel O Stram c, Loïc Le Marchand g, Lynne R Wilkens g, Beate Ritz h, Anna H Wu c
PMCID: PMC10908249  NIHMSID: NIHMS1967181  PMID: 35063793

Abstract

Background:

Within the Multiethnic Cohort (MEC), we examined the association between air pollution and mortality among African American, European American, Japanese American, and Latina American women diagnosed with breast cancer.

Methods:

We used a land use regression (LUR) model and kriging interpolation to estimate nitrogen oxides (NOx, NO2) and particulate matter (PM2.5, PM10) exposures for 3,089 breast cancer cases in the MEC, who were diagnosed from 1993 through 2013 and resided largely in Los Angeles County, California. Cox proportional hazards models were used to examine the association of time-varying air pollutants with all-cause, breast cancer, cardiovascular disease (CVD), and non-breast cancer/non-CVD mortality, accounting for key covariates.

Results:

We identified 1,125 deaths from all causes (474 breast cancer, 272 CVD, 379 non-breast cancer/non-CVD deaths) among the 3,089 breast cancer cases with 8.1 years of average follow-up. LUR and kriged NOX (per 50 ppb) and NO2 (per 20 ppb), PM2.5 (per 10 μg/m3), and PM10 (per 10 μg/m3) were positively associated with risks of all-cause (Hazard Ratio (HR) range = 1.13–1.25), breast cancer (HR range = 1.19–1.45), and CVD mortality (HR range = 1.37–1.60). Associations were statistically significant for LUR NOX and CVD mortality (HR = 1.60; 95% CI: 1.08–2.37) and kriged NO2 and breast cancer mortality (HR = 1.45; 95% CI 1.02–2.07). Gaseous and PM pollutants were positively associated with breast cancer mortality across racial/ethnic group.

Conclusion:

In this study, air pollutants have a harmful impact on breast cancer survival. Additional studies should evaluate potential confounding by socioeconomic factors. These data support maintaining clean air laws to improve survival for women with breast cancer.

1. Introduction

Breast cancer is the most commonly diagnosed non-skin cancer among U.S. women. >3.8 million breast cancer survivors were estimated in the U.S. in 2019 and this number of survivors is projected to increase to 4.9 million by 2030 (Miller et al., 2019). Prognostic factors for survival following breast cancer diagnosis include stage and other tumor characteristics, treatment factors, co-morbidities, sociodemographic factors, as well as modifiable lifestyle factors such as smoking, obesity, and physical inactivity (Hellmann et al., 2010; Lu et al., 20152015; Kwan et al., 2014). It is a public health priority to identify modifiable factors that improve outcome among this large and growing number of breast cancer survivors.

Air pollutants have been well documented to impact adversely numerous health outcomes, including mortality, particularly from cardiorespiratory diseases (Kelly and Fussell, 2015; Hoek et al., 2013). In a large US study, exposure to particulate matter (PM) <2.5 μm in diameter was estimated to be responsible for over 15,000 female deaths from 1999 through 2015, and having greatest effect of loss in life expectancy in some southern states and Los Angeles, California (Bennett et al., 2019). It has been estimated that >40% of people in the US live in counties with unhealthy air quality, placing them at risk for adverse health outcomes, particularly vulnerable groups such as those with chronic conditions (American Lung Association Association, 2019).

To date, two U.S. studies have examined the association between air pollution and survival following breast cancer diagnosis (DuPré et al., 2019; Hu et al., 2013). In a report of 255,128 breast cancer cases based on data from the Surveillance Epidemiology End Results (SEER) program, estimates of particulate matter<10 μm in diameter (PM10) and PM2.5 were significantly associated with increased breast cancer mortality with stronger associations seen for localized disease (Hu et al., 2013). In the Nurses’ Health Study (NHS) with 8,936 breast cancer cases, estimates of PM10 and PM2.5 based on 2-year averages were not associated with breast cancer mortality, although a statistically significant association was observed with PM2.5 among cases with stage I disease (DuPré et al., 2019). An Italian study (Tagliabue et al., 2016) reported PM2.5 based on a median of 3 years around diagnosis was associated with breast cancer mortality and a study in China (Huo et al., 2015) reported suggestive findings for an annual PM estimate and all-cause mortality among estrogen receptor (ER) positive breast cancers. However, these latter two studies did not consider lifestyle factors (Huo et al., 2015; Troeschel et al., 20192019). There is a need for additional studies of air pollution and breast cancer survival with adjustment for relevant individual-level covariates. In addition, investigations are warranted to evaluate whether certain racial/ethnic and/or socioeconomic groups, who often reside in areas with higher levels of air pollution (Turner et al., 2011; Wang et al., 2020), experience different mortality hazards in relation to air pollutant exposure.

Thus, we conducted a prospective study of traffic-related air pollution exposures and mortality among California female participants of the Multiethnic Cohort Study (MEC) with a breast cancer diagnosis from 1993 through 2013.

2. Methods

2.1. Study subjects

The MEC is a large population-based prospective cohort of US adults (Kolonel et al., 2000). Briefly, from 1993 through 1996, 96,810 men and 118,441 women aged 45–75 years largely from five self-reported racial/ethnic groups residing in Hawaii (HI) or CA (primarily Los Angeles County) were enrolled. Participants completed a baseline questionnaire that surveyed demographics, anthropometrics, reproductive history, and other lifestyle factors. Participants were followed for diagnosis of incident invasive breast cancer through routine linkage with the HI and CA cancer SEER registries, and for vital status through linkages to the National Death Index and state death certificate files that provide primary cause of death based on International Classification of Disease (ICD)-9 and ICD-10 codes. Clinicopathologic and treatment factors were obtained from the cancer registries. For this study, eligible female CA MEC participants with primary invasive breast cancer (ICD-O-3 C500-C509) were those who completed a baseline questionnaire with valid addresses across the study period (n = 3,113). We excluded cases with implausible or insufficient address data (n = 44), leaving 3,089 breast cancer cases for analyses. This cohort was followed from the date of diagnosis (1993–2013) to the date of death or December 31, 2013 (study end date), whichever came earlier. The institutional review boards of the University of Hawaii, University of Southern California, and University of California, San Francisco approved the study protocol.

2.2. Participant characteristics

Participant characteristics that we evaluated were age at diagnosis, race/ethnicity and baseline variables (median = 7.3, Q1 = 3.3; Q3 = 12.6 years between cohort entry and death) including marital status (married, single, divorced/widowed), body mass index (BMI) (under-weight (<18.5 kg/m2), normal weight (18.5–24.9 kg/m2), overweight (25–29.9 kg/m2), and obese (≥30 kg/m2)), smoking status (never, former, current smoker), alcohol intake (non-drinker (0 g alcohol/day), drinker (>0 g alcohol/day)), diabetes (yes, no), cardiovascular disease (CVD) (coronary heart disease and/or stroke, hypertension and/or hypertension medications, none), and age at first live birth (no children, ≤20, 21–30, >30 years). Clinicopathologic information included stage at diagnosis (localized, regional, distant), grade (I, II, and III & IV), histology (ductal, lobular, other, inflammatory breast cancer), ER/progesterone receptor (PR) status (ER + PR+, ER + PR-, ER-PR+, ER-PR-), and tumor size (<1cm, 1-<5cm, ≥5cm). First course of treatment included surgery (no, lumpectomy, mastectomy), chemotherapy (no, yes), hormone therapy (no, yes), and radiation (no, yes). The frequency of missing data for variables such as BMI, smoking, age at first live birth, and stage was low (≤2.9%).

2.3. Address history, geocoding, and neighborhood socioeconomic status

The MEC actively maintains a residential history of address locations on all participants based on mailings, linkages to secondary data sources, and direct communication from study participants. For the 3,089 CA MEC breast cancer cases, there were 4,305 residential addresses (74.5% non-movers, 22.4% 1–2 moves, 3.1% 3 + moves). Residential addresses were geocoded to latitude and longitude coordinates using address or street locators. Geocoded addresses were linked to 1990 (1993–1996 addresses), 2000 (1997–2005 addresses), and 2010 (2006–2013 addresses) US Census block groups. An index measure of nSES was based on principal component analysis of seven census-based indicators of SES: education, median household income, percent living 200% below the poverty level, percent blue-collar workers, percent without a job among those older than 15 years in workforce, median rent, and median house value (Yost et al., 2001; Yang et al., 2014). This nSES index was assigned to participants’ address at diagnosis and at time of the event of death or censoring and was categorized into quintiles based on the nSES distribution of Los Angeles County block groups for each decennial census year.

2.4. Air pollution exposure assessment

We used established air pollution assessment approaches (Cheng et al., 2020) based on linkage of geocoded residential addresses with latitude/longitude coordinates as the geographic unit to estimate traffic-related air pollutant exposures. In brief, a land use regression (LUR) model estimated NOx and NO2 exposures from regional and local sources based on air monitoring data from spatially dense air monitoring campaigns (2006–2007) and incorporated spatial data on land use and traffic characteristics; for temporal adjustment, monthly scaling factors were applied based on routinely collected air monitoring data nearest to the participant’s residence (Jerrett et al., 2005). Empirical Bayesian kriging interpolation was used to estimate largely regional exposures for NOx, NO2, PM10, and PM2.5 (Laurent et al., 2016). Measured concentrations of NOx, NO2, PM10 (1993–2013) and PM2.5 (2000–2013) were obtained from routinely collected air monitoring data from the US EPA. PM2.5 concentrations for 1993–1999 were estimated from a spatiotemporal model (Li et al., 2017). A correlation matrix of pollutants is presented in Supplemental Table 1 (r = 0.59–0.98).

2.5. Statistical analysis

We used Cox regression with time-varying exposure variables to examine the hazards of four mortality endpoints: all-cause, breast cancer, CVD, and non-breast/non-CVD mortality in relation to air pollution exposure. The Cox regression model used age in months as the time variable and defined a series of risk sets based on the age of death (months) of each event (index death). Each risk set consisted of all breast cancer cases who died at the age of the index death or remained alive and uncensored at that same age. For each member of each risk set (including the index death), we used her residential history to compute an average air pollutant exposure from time of diagnosis (month/year) to time of death of the index case in each risk set.

The Cox regression model used age of breast cancer diagnosis as the strata variable, and adjusted for demographics, lifestyle factors at baseline, clinicopathologic and treatment factors at diagnosis, and nSES at diagnosis and at death/censoring. Supplemental Table 2 presents the associations of covariates and all-cause mortality among breast cancer cases in the LUR NOx model. For Hazard ratios (HR) and 95% Confidence Intervals (CI) of cause specific deaths, we used a competing risk model where the at-risk denominator included living participants up until censoring at the time of death from other causes. HRs and 95% CI sfor common fixed size increases in air pollutants were calculated. For NOx, we used 50 ppb, which was close to the interquartile range (IQRs) of the LUR (41.1 ppb) and kriged (42.5 ppb) estimates. For NO2, we used 20 ppb, which was close to the IQR for LUR (13.3 ppb) and kriged (14.1 ppb) estimates. For PM10 and PM2.5, we used 10 μg/m3, which was between the IQR of kriged PM10 (12.5 μg/m3) and PM2.5 (6.4 μg/m3). We checked the proportional hazards assumption for each pollutant and found no violation. Subgroup analyses were conducted for race/ethnicity, hormone receptor-positive (ER + or PR + denoted as ER+/PR +) and hormone receptor-negative (ER- and PR- denoted as ER-PR-) breast tumors, first course of treatment, stage, nSES at diagnosis, pre-existing CVD (CVD status at baseline), and moving status. We assessed heterogeneity in associations using a global test of interaction.

To understand differences in air pollutant-mortality associations by nSES, we plotted the linear trends in HR for mortality across NO2 levels categorized into 20 quantiles by low (Q1–Q3) and high (Q4–Q5) nSES at diagnosis. The reference was the combination of the lowest NO2 level (first quantile) and high nSES. HRs were adjusted for all covariates described above.

All P values are two-sided with a statistical significance level of 0.05. Analyses were performed using SAS 9.2 statistical software (SAS Institute, Cary, NC).

3. Results

The study population consisted of 3,089 breast cancer cases (38% African American, 19% European American, 12% Japanese American, 31% Latina American, and 0.4% Native Hawiian) with racial/ethnic differences in age at diagnosis, marital status, obesity, smoking, alcohol intake, comorbidities, and age at first birth (Table 1). Regional/distant disease was higher among African Americans (33.6%) than other racial/ethnic groups (range: 22.4%–32.5%). The proportion of ER-PR- tumors was highest in African American (19.2%) and lowest in European American (10.8%) breast cancer cases. African American (27.0%) and Latina American breast cancer cases (20.3%) were more likely to live in the lowest SES neighborhoods at diagnosis than European American (7.7%) and Japanese American (4.3%) breast cancer cases. African American and Latina American breast cancer cases had higher average NOX exposures in comparison to Japanese American and European American breast cancer cases (Supplemental Table 3).

Table 1.

Study characteristics of 3,089 women diagnosed with breast cancer in CA MEC (1993–2013).

All (n = 3089) African Americans (n = 1,177) European Americans (n = 573) Japanese Americans (n = 370) Latina Americans (n = 958)
n % n % n % n % n %
Race/Ethnicitya African American 1177 38.1% 1177 100% – – – – – –
European American 573 18.6% – – 573 100% – – – –
Japanese American 370 12.0% – – – – 370 100% – –
Latina American 958 31.0% – – – – – – 958 100%
Native Hawaiian 11 0.4% – – – – – – – –
Age at diagnosis, years 45–49 35 1.1% 14 1.2% 5 0.9% 4 1.1% 12 1.3%
50–54 144 4.7% 68 5.8% 20 3.5% 24 6.5% 32 3.3%
55–59 289 9.4% 102 8.7% 57 10.0% 31 8.4% 97 10.1%
60–64 443 14.3% 172 14.6% 64 11.2% 42 11.4% 159 16.6%
65–69 599 19.4% 197 16.7% 123 21.5% 75 20.3% 201 21.0%
70–74 605 19.6% 208 17.7% 113 19.7% 80 21.6% 204 21.3%
75+ 974 31.5% 416 35.3% 191 33.3% 114 30.8% 253 26.4%
Marital statusa,b Married 1614 52.3% 438 37.2% 352 61.4% 271 73.2% 547 57.1%
Single 205 6.6% 63 5.4% 27 4.7% 31 8.4% 84 8.8%
Divorced/Widowed 1229 39.8% 649 55.1% 188 32.8% 68 18.4% 319 33.3%
BMIa,b Underweight 35 1.1% 9 0.8% 8 1.4% 12 3.2% 6 0.6%
Normal 1015 32.9% 258 21.9% 255 44.5% 233 63.0% 268 28.0%
Overweight 1094 35.4% 421 35.8% 176 30.7% 105 28.4% 385 40.2%
Obese 913 29.6% 461 39.2% 133 23.2% 20 5.4% 296 30.9%
Smoking statusa,b Never smoker 1655 53.6% 524 44.5% 267 46.6% 250 67.6% 610 63.7%
Former smoker 921 29.8% 390 33.1% 216 37.7% 98 26.5% 213 22.2%
Current smoker 446 14.4% 236 20.1% 84 14.7% 20 5.4% 103 10.8%
Alcohol intakea,b Non-drinker 1747 56.6% 696 59.1% 232 40.5% 251 67.8% 561 58.6%
Drinker 1236 40.0% 447 38.0% 318 55.5% 105 28.4% 362 37.8%
Diabetesa No 2749 89.0% 1028 87.3% 535 93.4% 344 93.0% 833 87.0%
Yes 340 11.0% 149 12.7% 38 6.6% 26 7.0% 125 13.1%
CVDa Coronary heart disease, stroke 300 9.7% 148 12.6% 42 7.3% 16 4.3% 94 9.8%
Hypertension, hypertension medications 1137 36.8% 562 47.8% 169 29.5% 126 34.1% 274 28.6%
None 1652 53.5% 467 39.7% 362 63.2% 228 61.6% 590 61.6%
Age at first live birtha,b Nulliparous 420 13.6% 156 13.3% 92 16.1% 75 20.3% 97 10.1%
15–20y 939 30.4% 489 41.6% 124 21.6% 20 5.4% 304 31.7%
21–30y 1421 46.0% 438 37.2% 296 51.7% 215 58.1% 463 48.3%
>30y 221 7.2% 52 4.4% 54 9.4% 49 13.2% 66 6.9%
Stagec Localized 2070 67.0% 758 64.4% 385 67.2% 285 77.0% 634 66.2%
Regional 865 28.0% 345 29.3% 166 29.0% 77 20.8% 274 28.6%
Distant 106 3.4% 50 4.3% 20 3.5% 6 1.6% 30 3.1%
Gradec Grade I 610 19.8% 210 17.8% 136 23.7% 97 26.2% 166 17.3%
Grade II 1183 38.3% 385 32.7% 241 42.1% 157 42.4% 394 41.1%
Grade III & IV 984 31.9% 434 36.9% 149 26.0% 95 25.7% 302 31.5%
Histologyc Ductal 2191 70.9% 854 72.6% 374 65.3% 274 74.1% 678 70.8%
Lobular 529 17.1% 181 15.4% 129 22.5% 58 15.7% 161 16.8%
Other 336 10.9% 132 11.2% 62 10.8% 36 9.7% 106 11.1%
Inflammatory breast cancer 33 1.1% 10 0.9% 8 1.4% 2 0.5% 13 1.4%
Estrogen/Progesterone Receptor statusc ER + PR+ 1714 55.5% 609 51.7% 356 62.1% 236 63.8% 504 52.6%
ER + PR- 339 11.0% 127 10.8% 72 12.6% 39 10.5% 100 10.4%
ER-PR+ 50 1.6% 22 1.9% 8 1.4% 4 1.1% 16 1.7%
ER-PR- 502 16.3% 226 19.2% 62 10.8% 48 13.0% 165 17.2%
Tumor sizec <1 cm 509 16.5% 154 13.1% 119 20.8% 85 23.0% 148 15.5%
1-<5 cm 2136 69.2% 835 70.9% 385 67.2% 248 67.0% 660 68.9%
≥5 cm 264 8.6% 110 9.4% 43 7.5% 27 7.3% 84 8.8%
Surgery typec No surgery 186 6.0% 96 8.2% 27 4.7% 14 3.8% 49 5.1%
Lumpectomy 821 26.6% 321 27.3% 145 25.3% 92 24.9% 261 27.2%
Mastectomy 1681 54.4% 599 50.9% 335 58.5% 217 58.7% 523 54.6%
Chemotherapyc No 2329 75.4% 881 74.9% 442 77.1% 291 78.7% 708 73.9%
Yes 760 24.6% 296 25.2% 131 22.9% 79 21.4% 250 26.1%
Hormone therapyc No 2031 65.8% 777 66.0% 361 63.0% 240 64.9% 647 67.5%
Yes 1058 34.3% 400 34.0% 212 37.0% 130 35.1% 311 32.5%
Radiationc No 1888 61.1% 765 65.0% 309 53.9% 199 53.8% 610 63.7%
Yes 1201 38.9% 412 35.0% 264 46.1% 171 46.2% 348 36.3%
Neighborhood SESa,c Quintile 1-low 572 18.5% 318 27.0% 44 7.7% 16 4.3% 194 20.3%
Quintile 2 668 21.6% 315 26.8% 73 12.7% 29 7.8% 246 25.7%
Quintile 3 744 24.1% 277 23.5% 131 22.9% 75 20.3% 259 27.0%
Quintile 4 613 19.8% 154 13.1% 159 27.8% 128 34.6% 169 17.6%
Quintile 5-high 459 14.9% 98 8.3% 162 28.3% 120 32.4% 78 8.1%
Open in a new tab
a

At baseline

b

Numbers may not total to 100% due to missing

c

At Diagnosis

LUR and kriged NOx (per 50 ppb) and NO2 (per 20 ppb), and PM2.5 and PM10 (per 10 μg/m3) were positively associated with all-cause (HR range = 1.13–1.25; p value range = 0.05–0.25) and breast cancer mortalities (HR range = 1.19–1.45; p value range = 0.04–0.29) (Table 2). For breast cancer mortality, a statistically significant increased risk was observed for kriged NO2 (HR = 1.45; 95% CI: 1.02–2.07). For CVD mortality, a statistically significant increased risk was observed for LUR NOx (HR = 1.60; 95% CI: 1.08–2.37).

Table 2.

Associations of gaseous and particulate matter air pollutants and risk of death following breast cancer diagnosis, CA MEC breast cancer cases, 1993–2013a.

All-cause mortality Breast cancer mortality CVD mortality Non-breast cancer and non-CVD mortalityc
Deaths HR 95% CI p-value Deaths HR 95% CI p-value Deaths HR 95% CI p-value Deaths HR 95% CI p-value
LUR NOxb 981 1.16 (0.96–1.40) 0.12 418 1.25 (0.93–1.66) 0.13 240 1.60 (1.08–2.37) 0.02 323 0.91 (0.64–1.30) 0.61
NO2b 1005 1.14 (0.91–1.43) 0.25 428 1.34 (0.95–1.90) 0.09 245 1.49 (0.92–2.40) 0.11 332 0.81 (0.53–1.22) 0.31
Kriging NOxb 1095 1.16 (0.96–1.39) 0.13 461 1.27 (0.96–1.68) 0.09 269 1.37 (0.93–2.02) 0.12 365 0.93 (0.67–1.29) 0.65
NO2b 1111 1.25 (1.00–1.57) 0.05 468 1.45 (1.02–2.07) 0.04 269 1.58 (0.99–2.51) 0.06 374 0.94 (0.64–1.38) 0.75
PM2.5b 1109 1.17 (0.95–1.44) 0.13 465 1.19 (0.86–1.64) 0.29 269 1.44 (0.95–2.17) 0.08 375 0.94 (0.66–1.34) 0.72
PM10b 1112 1.13 (1.00–1.29) 0.06 468 1.20 (0.98–1.48) 0.09 269 1.25 (0.97–1.62) 0.09 375 0.96 (0.77–1.21) 0.75
Open in a new tab
a

Adjusted for race/ethnicity, marital status, ERPR status, stage, grade, histology, tumor size, chemotherapy, hormone treatment, radiation, surgery, nSES at diagnosis and current nSES, BMI, smoking, alcohol, diabetes, CVD risk, and age at first live birth; and stratified by age at diagnosis

b

LUR and kriging per 50 ppb NOX; 20 ppb Kriging and LUR NO2; Kriging per 10ug/m3 PM10 and PM2.5

c

Distribution of cause of deaths: other cancer site than breast cancer (42%), diabetes (9%), chronic lower respiratory disease (9%), Alzheimer’s disease (8%), mental/behavioral disease (5%), pneumonia (5%), kidney disease (3%), liver disease (3%), infectious/parasitic disease (3%), other causes of deaths (5%), unknown cause (8%)

For all-cause mortality, similar patterns of increased risk were seen for all pollutants among African American and European American breast cancer cases; results reached formal statistical significance for LUR NOx and PM10 among African American breast cancer cases and kriged NOx and NO2 among European American breast cancer cases (Table 3). For breast cancer mortality, consistent patterns of increased risks were seen for all pollutants across race/ethnicity with a statistically significantly increased risk for kriged NOX and NO2 among European American breast cancer cases. For CVD mortality, 1.6–3.6-fold increased risks of death were seen for all pollutants among African American breast cancer cases (p value range = 0.03 to < 0.0001). For non-breast cancer/non-CVD mortality, increased risk associated with air pollutants were observed among European American breast cancer cases (HR range = 1.29–2.69) with wide 95% CIs. Almost all LUR and kriged pollutants displayed relatively larger HRs among ER-PR- breast cancer in comparison to ER+/PR + disease for all-cause (HR range = 1.21–1.81 vs. HR range = 1.03–1.27, respectively) and breast cancer mortalities (HR range = 1.47–2.63 vs. HR range = 1.05–1.53, respectively) (Table 4). Yet, there was no formal statistical evidence of heterogeneity in associations by ER/PR status.

Table 3.

Associations of gaseous and particulate matter air pollutants and risk of death following breast cancer diagnosis among CA MEC breast cancer cases by race/ethnicity, 1993–2013a,b.

All-cause mortality Breast cancer mortality CVD mortality Non-breast cancer and non-CVD mortality
Deaths HR 95% CI p-value Deaths HR 95% CI p-value Deaths HR 95% CI p-value Deaths HR 95% CI p-value
LUR NOxc African Americans 419 1.42 (1.04–1.94) 0.03 193 1.41 (0.88–2.25) 0.15 106 3.62 (1.91–6.86) <0.0001 120 0.83 (0.40–1.72) 0.62
European Americans 198 1.35 (0.80–2.30) 0.26 66 1.94 (0.63–6.02) 0.25 52 0.51 (0.09–2.72) 0.43 80 2.66 (1.0–7.07) 0.05
Japanese Americans 80 1.91 (0.63–5.83) 0.25 – – – – – – – – – – – –
Latina Americans 282 1.05 (0.71–1.54) 0.82 126 1.46 (0.79–2.71) 0.23 61 1.14 (0.43–2.97) 0.79 95 0.51 (0.25–1.05) 0.07
Pheterogenerity 0.46 0.82 0.02 0.02
NO2c African Americans 437 1.33 (0.91–1.95) 0.14 199 1.73 (0.97–3.08) 0.07 110 3.34 (1.50–7.44) 0.003 128 0.53 (0.23–1.21) 0.13
European Americans 203 1.30 (0.72–2.34) 0.39 69 2.23 (0.66–7.53) 0.20 53 0.88 (0.16–5.00) 0.89 81 1.71 (0.60–4.87) 0.32
Japanese Americans 80 3.18 (0.90–11.20) 0.07 – – – – – – – – – – – –
Latina Americans 283 1.02 (0.63–1.65) 0.92 127 1.17 (0.56–2.44) 0.68 61 1.00 (0.30–3.30) 1.00 95 0.68 (0.28–1.64) 0.39
Pheterogenerity 0.25 0.50 0.14 0.21
Kriging NOxc African Americans 491 1.18 (0.90–1.56) 0.24 220 1.24 (0.82–1.88) 0.32 124 2.16 (1.21–3.84) 0.01 147 0.74 (0.42–1.29) 0.28
European Americans 224 1.68 (1.01–2.80) 0.04 76 3.98 (1.31–12.04) 0.01 60 0.88 (0.23–3.36) 0.86 88 2.25 (0.86–5.85) 0.10
Japanese Americans 82 0.92 (0.26–3.28) 0.89 – – – – – – – – – – – –
Latina Americans 296 1.14 (0.76–1.72) 0.52 131 1.67 (0.87–3.19) 0.12 63 0.96 (0.33–2.79) 0.93 102 0.74 (0.36–1.51) 0.41
Pheterogenerity 0.91 0.42 0.09 0.73
NO2c African Americans 506 1.30 (0.91–1.86) 0.15 227 1.52 (0.87–2.65) 0.14 124 3.39 (1.59–7.23) 0.002 155 0.56 (0.28–1.10) 0.09
European Americans 224 1.87 (1.05–3.34) 0.03 76 4.90 (1.36–17.71) 0.02 60 0.85 (0.21–3.43) 0.82 88 2.69 (0.94–7.73) 0.07
Japanese Americans 83 0.88 (0.24–3.26) 0.84 – – – – – – – – – – – –
Latina Americans 296 1.19 (0.73–1.95) 0.49 131 1.61 (0.75–3.47) 0.22 63 1.05 (0.29–3.85) 0.94 102 0.85 (0.35–2.03) 0.71
Pheterogenerity 0.59 0.25 0.12 0.05
PM2.5c African Americans 504 1.35 (0.99–1.85) 0.06 225 1.33 (0.80–2.20) 0.27 124 2.10 (1.13–3.91) 0.02 155 0.92 (0.51–1.65) 0.78
European Americans 225 1.33 (0.77–2.30) 0.30 76 2.21 (0.63–7.82) 0.22 60 0.78 (0.20–3.02) 0.72 89 1.79 (0.69–4.65) 0.23
Japanese Americans 83 0.57 (0.15–2.20) 0.42 – – – – – – – – – – – –
Latina Americans 295 0.94 (0.60–1.47) 0.78 130 1.17 (0.59–2.33) 0.66 63 0.93 (0.29–3.02) 0.91 102 0.62 (0.28–1.36) 0.23
Pheterogenerity 0.23 0.87 0.11 0.63
PM10c African Americans 506 1.25 (1.03–1.52) 0.02 227 1.33 (0.97–1.83) 0.08 124 1.55 (1.04–2.29) 0.03 155 0.97 (0.68–1.37) 0.85
European Americans 225 1.23 (0.83–1.81) 0.31 76 1.51 (0.63–3.64) 0.36 60 1.05 (0.38–2.93) 0.93 89 1.29 (0.67–2.46) 0.45
Japanese Americans 83 0.85 (0.34–2.12) 0.73 – – – – – – – – – – – –
Latina Americans 296 0.98 (0.73–1.30) 0.87 131 1.11 (0.72–1.73) 0.64 63 1.24 (0.54–2.82) 0.61 102 0.79 (0.47–1.33) 0.36
Pheterogenerity 0.21 0.52 0.34 0.66
Open in a new tab
a

African American (n = 1117), European American (n = 573), Japanese American (n = 370), and Latina American (n = 958) breast cancer cases. Native Hawaiian (n = 11) breast cancer cases were not included.

b

Adjusted for marital status, ERPR status, stage, grade, histology, tumor size, chemotherapy, hormone treatment, radiation, surgery, nSES at diagnosis and current nSES, BMI, smoking, alcohol, diabetes, CVD risk, and age at first live birth; age at diagnosis as a stratum variable

c

LUR and kriging per 50 ppb NOX; 20 ppb Kriging and LUR NO2; Kriging per 10ug/m3 PM10 and PM2.5

Table 4.

Associations of gaseous and particulate matter air pollutants and risk of death following breast cancer diagnosis among CA MEC breast cancer cases by ERPR status, 1993–2013a.

All-cause mortality Breast cancer mortality CVD mortality Non-breast cancer and non-CVD mortality
Deaths HR 95% CI p-value Deaths HR 95% CI p-value Deaths HR 95% CI p-value Deaths HR 95% CI p-value
LUR NOxb ER+/PR+ 569 1.27 (0.97–1.66) 0.08 214 1.53 (0.98–2.39) 0.06 144 1.35 (0.78–2.33) 0.28 211 1.05 (0.67–1.64) 0.85
ER-PR- 189 1.20 (0.72–1.99) 0.48 106 1.58 (0.81–3.09) 0.18 – – – – 48 1.30 (0.32–5.28) 0.71
Pheterogenerity 0.84 0.92 – 0.74
NO2b ER+/PR+ 584 1.20 (0.87–1.64) 0.26 220 1.34 (0.79–2.26) 0.28 148 1.25 (0.65–2.42) 0.51 216 1.10 (0.65–1.85) 0.73
ER-PR- 196 1.21 (0.67–2.15) 0.53 109 1.95 (0.89–4.28) 0.10 – – – – 51 1.67 (0.31–9.03) 0.55
Pheterogenerity 0.98 0.41 – 0.64
Kriging NOxb ER+/PR+ 647 1.14 (0.89–1.47) 0.31 238 1.31 (0.86–2.02) 0.21 166 1.33 (0.79–2.24) 0.29 243 0.98 (0.65–1.49) 0.93
ER-PR- 213 1.40 (0.89–2.20) 0.15 120 1.76 (0.93–3.31) 0.08 41 2.64 (0.42–16.76) 0.30 52 0.87 (0.27–2.78) 0.81
Pheterogenerity 0.40 0.40 0.38 0.86
N02b ER+/PR+ 659 1.08 (0.80–1.45) 0.63 243 1.28 (0.76–2.14) 0.35 166 1.36 (0.73–2.55) 0.33 250 0.89 (0.55–1.44) 0.64
ER-PR- 215 1.81 (1.04–3.15) 0.04 122 2.63 (1.19–5.83) 0.02 41 1.79 (0.21–15.2) 0.59 52 1.49 (0.33–6.83) 0.61
Pheterogenerity 0.11 0.14 0.81 0.53
PM2.5b ER+/PR+ 657 1.03 (0.78–1.36) 0.84 241 1.05 (0.65–1.72) 0.83 166 1.12 (0.64–1.97) 0.69 250 0.98 (0.63–1.51) 0.91
ER-PR- 215 1.41 (0.86–2.30) 0.17 121 1.76 (0.88–3.52) 0.11 41 1.30 (0.20–8.43) 0.78 53 1.22 (0.33–4.58) 0.77
Pheterogenerity 0.20 0.18 0.72 0.72
PM10b ER+/PR+ 659 1.09 (0.92–1.29) 0.33 243 1.12 (0.83–1.52) 0.46 166 1.12 (0.80–1.58) 0.51 250 1.07 (0.81–1.40) 0.65
ER-PR- 216 1.32 (0.96–1.81) 0.09 122 1.47 (0.95–2.28) 0.09 41 0.65 (0.20–2.08) 0.46 53 1.37 (0.57–3.31) 0.48
Pheterogenerity 0.15 0.19 0.39 0.63
Open in a new tab
a

Adjusted for race/ethnicity, marital status, stage, grade, histology, tumor size, chemotherapy, hormone treatment, radiation, surgery, nSES at diagnosis and current nSES, BMI, smoking, alcohol, diabetes, CVD risk, and age at first live birth; age at diagnosis as a stratum variable

b

Kriging and LUR per 50 ppb NOX; 20 ppb Kriging and LUR NO2; Kriging per PM10 and PM2.5

Risk of CVD mortality associated with air pollutants was higher among low than high nSES cases (Table 5; p heterogeneity > 05). However, LUR NO2 was associated with increased risk of all-cause, breast cancer, and non-breast cancer/non-CVD mortalities among high nSES breast cancer cases, but either inversely or not associated with these outcomes among low nSES cases (p heterogeneity ≤ 05). Supplemental Figure 1 shows that at low levels of LUR NO2, the HRs for all-cause and breast cancer mortalities were higher among low nSES cases. In contrast, at higher levels of LUR NO2, the HRs for these outcomes were higher for high nSES cases. That is, while starting at lower risk, the slope of increase in risk was steeper among high nSES cases with increasing LUR NO2.

Table 5.

Associations of gaseous and particulate matter air pollutants and risk of death following breast cancer diagnosis among CA MEC breast cancer cases by nSES at diagnosis, 1993–2013a.

All-cause mortality Breast cancer mortality CVD mortality Non-breast cancer and non-CVD mortality
Deaths HR 95% CI p-value Deaths HR 95% CI p-value Deaths HR 95% CI p-value Deaths HR 95% CI p-value
LUR NOxb low nSES (Q1–Q3) 681 1.04 (0.82–1.31) 0.76 298 1.08 (0.75–1.55) 0.67 168 1.57 (0.99–2.50) 0.06 215 0.74 (0.47–1.16) 0.19
high nSES (Q4–Q5) 289 1.44 (0.97–2.13) 0.07 116 1.87 (0.98–3.55) 0.06 70 0.65 (0.23–1.81) 0.41 103 1.43 (0.72–2.83) 0.31
Pheterogenerity 0.15 0.14 0.11 0.09
N02b low nSES (Q1–Q3) 698 0.93 (0.69–1.24) 0.61 305 1.04 (0.66–1.63) 0.88 171 1.42 (0.78–2.59) 0.25 222 0.62 (0.36–1.08) 0.09
high nSES (Q4–Q5) 296 1.65 (1.08–2.53) 0.02 119 2.44 (1.19–5.00) 0.02 72 0.68 (0.24–1.92) 0.46 105 1.58 (0.74–3.35) 0.23
Pheterogenerity 0.03 0.05 0.22 0.05
Kriging NOxb low nSES (Q1–Q3) 758 1.12 (0.90–1.40) 0.32 326 1.23 (0.88–1.71) 0.23 188 1.48 (0.92–2.38) 0.10 244 0.85 (0.57–1.27) 0.43
high nSES (Q4–Q5) 325 1.31 (0.89–1.92) 0.17 131 1.48 (0.75–2.93) 0.26 78 0.90 (0.37–2.20) 0.82 116 1.37 (0.74–2.56) 0.32
Pheterogenerity 0.49 0.62 0.34 0.20
N02b low nSES (Q1–Q3) 771 1.15 (0.87–1.53) 0.33 331 1.42 (0.91–2.21) 0.12 188 1.69 (0.93–3.07) 0.08 252 0.77 (0.47–1.26) 0.30
high nSES (Q4–Q5) 328 1.49 (0.98–2.27) 0.07 133 1.46 (0.69–3.07) 0.32 78 0.92 (0.36–2.33) 0.86 117 1.86 (0.91–3.80) 0.09
Pheterogenerity 0.32 0.95 0.28 0.05
PM2.5b low nSES (Q1–Q3) 770 1.07 (0.83–1.38) 0.61 329 1.06 (0.71–1.58) 0.77 188 1.43 (0.85–2.42) 0.18 253 0.88 (0.56–1.37) 0.56
high nSES (Q4–Q5) 327 1.38 (0.93–2.04) 0.11 132 1.32 (0.67–2.60) 0.43 78 1.40 (0.59–3.35) 0.44 117 1.33 (0.67–2.65) 0.42
Pheterogenerity 0.29 0.59 0.97 0.32
PM10b low nSES (Q1–Q3) 772 1.06 (0.90–1.25) 0.50 331 1.16 (0.89–1.50) 0.28 188 1.11 (0.79–1.55) 0.55 253 0.93 (0.70–1.25) 0.64
high nSES (Q4–Q5) 328 1.30 (1.01–1.66) 0.04 133 1.37 (0.87–2.15) 0.18 78 1.39 (0.79–2.44) 0.25 117 1.16 (0.77–1.76) 0.48
Pheterogenerity 0.18 0.53 0.50 0.39
Open in a new tab
a

Adjusted for race/ethnicity, marital status, ERPR status, stage, grade, histology, tumor size, chemotherapy, hormone treatment, radiation, surgery, current nSES, BMI, smoking, alcohol, diabetes, CVD risk, and age at first live birth; age at diagnosis as a stratum variable

b

Kriging and LUR per 50 ppb NOX; 20 ppb Kriging and LUR NO2; Kriging per PM10 and PM2.5 PM10 and PM2.5

Largely similar patterns of associations were observed for localized and advanced disease (Table 6). For breast cancer cases with no pre-existing CVD, statistically significant increased risks for all-cause mortality were seen with LUR NOX (HR = 1.38; 95% CI: 1.04–1.82), kriged NOX (HR = 1.38; 95% CI: 1.03–1.85) and NO2 (HR = 1.52; 95% CI: 1.06–2.18) (Supplemental Table 4). While for cases with pre-existing CVD, kriged pollutants and LUR NO2 displayed larger HRs for CVD mortality than cases with no CVD. Analyses by breast cancer treatment showed that PM2.5 and PM10 were consistently associated with increased risk of all-cause, breast cancer, and CVD mortalities among cases who did not receive either chemotherapy or radiation (Supplemental Table 5). Results by moving status showed positive associations with all-cause, breast cancer, and CVD mortalities for all pollutants among non-movers while patterns were less consistent patterns among movers (Supplemental Table 6).

Table 6.

Associations of gaseous and particulate matter air pollutants and risk of death following breast cancer diagnosis among CA MEC breast cancer cases by stage of disease, 1993–2013a.

All-cause mortality Breast cancer mortality CVD mortality Non-breast cancer and non-CVD mortality
Deaths HR 95% CI p-value Deaths HR 95% CI p-value Deaths HR 95% CI p-value Deaths HR 95% CI p-value
LUR NOxb Localized 525 1.20 (0.91–1.59) 0.21 134 1.29 (0.77–2.15) 0.33 157 1.52 (0.87–2.64) 0.14 234 0.95 (0.61–1.48) 0.81
Advanced 428 1.03 (0.77–1.39) 0.82 269 1.05 (0.72–1.54) 0.80 78 1.25 (0.55–2.84) 0.59 81 0.93 (0.42–2.09) 0.87
Pheterogenerity 0.47 0.48 0.72 0.98
N02b Localized 536 1.11 (0.79–1.55) 0.56 138 1.20 (0.63–2.26) 0.58 159 1.35 (0.69–2.64) 0.38 239 0.88 (0.52–1.47) 0.62
Advanced 440 1.17 (0.82–1.68) 0.39 274 1.19 (0.74–1.90) 0.47 81 1.35 (0.53–3.41) 0.53 85 1.06 (0.41–2.75) 0.91
Pheterogenerity 0.81 0.99 1.00 0.74
Kriging NOxb Localized 589 1.21 (0.92–1.58) 0.17 150 1.55 (0.93–2.60) 0.09 177 1.40 (0.84–2.34) 0.19 262 0.95 (0.63–1.43) 0.80
Advanced 477 1.04 (0.78–1.39) 0.81 295 1.04 (0.72–1.51) 0.84 87 1.24 (0.54–2.84) 0.62 95 1.01 (0.52–1.98) 0.97
Pheterogenerity 0.48 0.21 0.76 0.91
N02b Localized 596 1.35 (0.98–1.86) 0.06 151 1.95 (1.02–3.75) 0.04 177 1.50 (0.83–2.71) 0.18 268 1.04 (0.65–1.68) 0.87
Advanced 485 1.14 (0.79–1.65) 0.48 301 1.11 (0.69–1.78) 0.66 87 1.22 (0.42–3.55) 0.72 97 1.06 (0.45–2.47) 0.90
Pheterogenerity 0.49 0.17 0.74 0.97
PM2.5b Localized 595 1.14 (0.85–1.52) 0.38 149 1.38 (0.76–2.51) 0.29 177 1.36 (0.80–2.31) 0.25 269 0.89 (0.58–1.36) 0.58
Advanced 484 1.12 (0.80–1.56) 0.51 300 0.97 (0.63–1.49) 0.88 87 1.15 (0.47–2.83) 0.76 97 1.67 (0.78–3.61) 0.19
Pheterogenerity 0.88 0.32 0.74 0.19
PM10b Localized 597 1.19 (0.99–1.43) 0.06 151 1.47 (1.01–2.13) 0.04 177 1.25 (0.90–1.73) 0.18 269 1.01 (0.77–1.33) 0.93
Advanced 485 1.08 (0.87–1.33) 0.49 301 1.00 (0.76–1.31) 0.99 87 0.95 (0.52–1.75) 0.87 97 1.21 (0.78–1.89) 0.39
Pheterogenerity 0.35 0.08 0.43 0.72
Open in a new tab
a

Adjusted for race/ethnicity, marital status, ERPR status, grade, histology, tumor size, chemotherapy, hormone treatment, radiation, surgery, nSES at diagnosis and current nSES, BMI, smoking, alcohol, diabetes, CVD risk, and age at first live birth; age at diagnosis as a stratum variable

b

LUR and kriging per 50 ppb NOX; 20 ppb Kriging and LUR NO2; Kriging per 10ug/m3 PM2.5 and PM10

4. Discussion

In this prospective study of female breast cancer cases in the MEC, NOX, NO2, PM2.5, and PM10 were positively associated with increased risks of all-cause, breast, and CVD mortalities. Consistently positive associations were observed across race/ethnicity for breast cancer mortality. In addition, almost all kriged and LUR pollutants displayed larger HRs for breast cancer mortality among ER-PR- breast cancer than ER+/PR + disease. Heterogeneity in LUR NO2 associations with all-cause and breast cancer mortalities by nSES were observed. Positive associations with all-cause, breast cancer and CVD mortalities were consistently observed among breast cancer cases who were non-movers. Overall, our findings provide new evidence that long-term air pollutant exposures adversely impact mortality outcomes for breast cancer survivors after adjustment for key covariates.

The positive associations observed for PM2.5 and PM10 (per 10 μg/m3; HR = 1.17 and 1.13, respectively) with all-cause mortality in the MEC support prior findings of the NHS based on time-varying spatiotemporal models (per 10 μg/m3 PM2.5 and PM10: HR = 1.12 and 1.09, respectively) (DuPré et al., 2019). In a California SEER-based study, county-level PM2.5 and PM10 were associated with increased risks of overall and breast cancer mortality for all breast cancer cases and those with localized disease (Hu et al., 2013). We also observed increased risks for breast cancer mortality with PM exposures but the 95% CIs included the null. The large SEER-based study (Hu et al., 2013) used county-level PM estimates based on a single address at diagnosis with limited individual-level covariates, while the MEC and NHS (DuPré et al., 2019) had fewer cases but used time-varying exposure estimates based on detailed residential histories and included individual-level covariate data.

The increased risks for CVD mortality associated with all air pollutants support the well-documented adverse health effects of air pollution on CVD (Brook et al., 2010). These overall CVD mortality associations were driven by associations observed among African American breast cancer cases, suggesting a particularly high risk for CVD mortality following their breast cancer diagnosis due to air pollutant exposures. Both chemotherapy and radiation treatment for breast cancer may increase the risk of CVD. In particular, anthracycline has known cardiotoxic effects while radiotherapy may increase CVD risk through injury to the cardiac muscle or the surrounding vasculature (Doyle et al., 2005; Accordino et al., 2014). Thus, it is surprising that the HRs for PM2.5 and PM10 with all-cause, breast cancer, and CVD mortalities were somewhat larger for women who did not receive chemotherapy and radiation than those who received chemotherapy and/or radiation, but the 95% CIs displayed large overlap between treatment strata.

We observed larger HRs for LUR NO2 with all-cause and breast cancer mortalities for breast cancer cases living in high versus low SES neighborhoods. However, at low levels of LUR NO2 larger HRs for all-cause and breast cancer mortalities were observed among low versus high nSES breast cancer cases, while at high levels of NO2, the relationship reversed in direction. As Supplemental Figure 1 illustrates, our findings suggest that high nSES breast cancer cases start with a lower mortality risk at low air pollution levels than low nSES cases but experience a steeper risk increase with increasing pollution that surpass low nSES cases at higher levels of LUR NO2. A similar pattern of lower baseline mortality risk and steeper increases by levels of air pollutant among the highest SES individuals was reported by a nationwide Danish study (Raaschou-Nielsen et al., 2020).

Air pollution is comprised a complex mixture of correlated gaseous pollutants and PM. We interpret the observed associations with the various air pollutants as reflecting a traffic-related air pollution mixture rather than any specific air pollutant. We did not conduct multi-pollutant modeling given the high degree of correlation between pollutants. Elevated effect estimates with LUR and kriging pollutants and all-cause mortality were consistently observed among African American and European American breast cancers with large numbers of deaths. Less consistent findings among Japanese American and Latina cases may be related to the smaller number of events overall for Japanese American cases and of non-breast cancer deaths among Latina American cases.

Air pollution includes various polycyclic aromatic hydrocarbons, metals, and benzene that are transported and metabolized throughout the body and have been linked to increased oxidative stress, inflammation, and epigenetic changes (Liu et al., 2019; Rider and Carlsten, 2019; Rao et al., 2018). Although the biological mechanisms by which air pollution may increase mortality among breast cancer cases are unclear, oxidative stress may adversely affect mortality through cell proliferation, genetic instability, and mutations (Kang and Hamasaki, 2003). Inflammation may also trigger the release of pro-inflammatory cytokines, leading to tissue and organ damage and death (Tsai et al., 2019; Li et al., 2019). Epigenetic changes, such as DNA methylation, resulting in the activation or silencing of key genes has been linked to mortality (Zhang et al., 2017).

To our knowledge, this is the first US study to evaluate associations between air pollution and mortality, overall and from different causes, among breast cancer cases across race/ethnicity, nSES, pre-existing CVD conditions, and first course of treatment. Study strengths include a diverse study population with regard to race/ethnicity and nSES. Our findings in Los Angeles County, an area that has experienced some of the highest air pollution levels in the US, provide important insights that may be particularly applicable to highly polluted megacities in developing and rapidly industrializing countries. In addition, we captured long-term air pollution exposures using detailed residential histories, an approach that possibly reduces exposure misclassification in comparison to studies limited to a single address at diagnosis. We also accounted for a large number of individual-level covariates and nSES.

Our study has limitations. While we accounted for nSES, using a construct that captured the domains of income, poverty, employment, and housing, we were unable to account for other unmeasured individual-level SES factors (e.g. insurance status). Information on air pollution exposures at non-residential locations and indoors, and details on treatment regimens and dose are lacking. Due to the smaller sample size in subgroup analyses, we acknowledge the imprecision in some of our effect estimates and the limited statistical power to detect associations for some outcomes (e.g. CVD mortality among Japanese Americans and ER-PR- breast cancers). We also acknowledge that a large number of comparisons were made with a possibility of false positive associations.

In conclusion, this study reports adverse effects of air pollution on mortality among breast cancer cases. While it remains to be determined whether breast cancer survivors are more susceptible to the adverse effects of air pollutants than the general population of women, maintaining stringent clean air laws serves as an actionable target to reduce mortality among the many US women with breast cancer. Future large studies of multiethnic and socioeconomically diverse populations are needed to corroborate and expand our findings.

Supplementary Material

Supp1
Supp2

Acknowledgements

This work was supported by Susan G. Komen (IIR13262718) for IC, JY, CT, JW, SMC, SSM, SLG, ASW, DS, BR, AW; Health Effects of Air Pollution Foundation (HEAPF014) for IC, JY CT, JW, SSM, AW; the National Cancer Institute (U01 CA164973) LRW, LLM; USC Norris Comprehensive Cancer Center (Core) Support (P30 CA014089) for AW; Environmental Exposures, Host Factors, and Human Disease (P30 ES0070480 for AW; California Air Resource Board (contract 04-323) for BR, National Cancer Institute (R21-CA094723). The authors are grateful to the study participants and research team of the MEC. We thank Diana Chingos who served as patient advocate on the study and Heather Rose for her administrative assistance.

Footnotes

Author Disclosures

None of the authors listed above have any conflicts of interest to disclose for this manuscript

Prior presentations

American Association of Cancer Research Conference on the Science of Cancer Health Disparities in Minorities and Medically Underserved, San Francisco 2019

CRediT authorship contribution statement

Iona Cheng: Conceptualization, Funding acquisition, Investigation. Juan Yang: Formal analysis, Investigation. Chiuchen Tseng: Formal analysis, Investigation. Jun Wu: Data curation, Investigation. Shannon M. Conroy: Investigation. Salma Shariff-Marco: Investigation. Scarlett Lin Gomez: Investigation. Alice S. Whittemore: Investigation. Daniel O. Stram: Investigation. Loïc Le Marchand: Investigation. Lynne R. Wilkens: Investigation. Beate Ritz: Data curation, Investigation. Anna H. Wu: Conceptualization, Funding acquisition, Investigation.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.envint.2022.107088.

References

  1. Miller KD, Nogueira L, Mariotto AB, Rowland JH, Yabroff KR, Alfano CM, Jemal A, Kramer JL, Siegel RL, 2019. Cancer treatment and survivorship statistics, 2019. CA Cancer J Clin 69 (5), 363–385. 10.3322/caac.21565. [DOI] [PubMed] [Google Scholar]
  2. Hellmann SS, Thygesen LC, Tolstrup JS, Grønbæk M, 2010. Modifiable risk factors and survival in women diagnosed with primary breast cancer: results from a prospective cohort study. Eur J Cancer Prev 19 (5), 366–373. [DOI] [PubMed] [Google Scholar]
  3. Lu Y, John EM, Sullivan-Halley J, Vigen C, Gomez SL, Kwan ML, Caan BJ, Lee VS, Roh JM, Shariff-Marco S, Keegan THM, Kurian AW, Monroe KR, Cheng I, Sposto R, Wu AH, Bernstein L, 2015. History of Recreational Physical Activity and Survival After Breast Cancer: The California Breast Cancer Survivorship Consortium. Am J Epidemiol 181 (12), 944–955. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Kwan ML, John EM, Caan BJ, Lee VS, Bernstein L, Cheng I, Gomez SL, Henderson BE, Keegan THM, Kurian AW, Lu Y, Monroe KR, Roh JM, Shariff-Marco S, Sposto R, Vigen C, Wu AH, 2014. Obesity and mortality after breast cancer by race/ethnicity: The California Breast Cancer Survivorship Consortium. Am J Epidemiol 179 (1), 95–111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Kelly FJ, Fussell JC, 2015. Air pollution and public health: emerging hazards and improved understanding of risk. Environ Geochem Health 37 (4), 631–649. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Hoek G, Krishnan RM, Beelen R, Peters A, Ostro B, Brunekreef B, Kaufman JD, 2013. Long-term air pollution exposure and cardio- respiratory mortality: a review. Environ Health 12 (1). 10.1186/1476-069X-12-43. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bennett JE, Tamura-Wicks H, Parks RM, et al. Particulate matter air pollution and national and county life expectancy loss in the USA: A spatiotemporal analysis. PLoS Med 2019;16(7):e1002856. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. DuPré NC, Hart JE, Holmes MD, Poole EM, James P, Kraft P, Laden F, Tamimi RM, 2019. Particulate Matter and Traffic-Related Exposures in Relation to Breast Cancer Survival. Cancer Epidemiol Biomarkers Prev 28 (4), 751–759. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Hu H, Dailey AB, Kan H, Xu X, 2013. The effect of atmospheric particulate matter on survival of breast cancer among US females. Breast Cancer Res Treat 139 (1), 217–226. [DOI] [PubMed] [Google Scholar]
  10. Tagliabue G, Borgini A, Tittarelli A, van Donkelaar A, Martin RV, Bertoldi M, Fabiano S, Maghini A, Codazzi T, Scaburri A, Favia I, Cau A, Barigelletti G, Tessandori R, Contiero P, 2016. Atmospheric fine particulate matter and breast cancer mortality: a population-based cohort study. BMJ Open 6 (11), e012580. 10.1136/bmjopen-2016-012580. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Huo Q, Cai C, Yang Q, 2015. Atmospheric particulate matter and breast cancer survival: estrogen receptor triggered? Tumour Biol 36 (5), 3191–3193. [DOI] [PubMed] [Google Scholar]
  12. Troeschel AN, Liu Y, Collin LJ, Bradshaw PT, Ward KC, Gogineni K, McCullough LE, 2019. Race differences in cardiovascular disease and breast cancer mortality among US women diagnosed with invasive breast cancer. Int J Epidemiol 48 (6), 1897–1905. 10.1093/ije/dyz108. [DOI] [PubMed] [Google Scholar]
  13. Turner MC, Krewski D, Pope CA, Chen Y, Gapstur SM, Thun MJ, 2011. Long-term ambient fine particulate matter air pollution and lung cancer in a large cohort of never-smokers. Am J Respir Crit Care Med 184 (12), 1374–1381. [DOI] [PubMed] [Google Scholar]
  14. Wang B, Eum K-D, Kazemiparkouhi F, Li C, Manjourides J, Pavlu V, Suh H, 2020. The impact of long-term PM2.5 exposure on specific causes of death: exposure-response curves and effect modification among 53 million U.S. Medicare beneficiaries. Environ Health 19 (1). 10.1186/s12940-020-00575-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Kolonel LN, Henderson BE, Hankin JH, Nomura AMY, Wilkens LR, Pike MC, Stram DO, Monroe KR, Earle ME, Nagamine FS, 2000. A multiethnic cohort in Hawaii and Los Angeles: baseline characteristics. Am J Epidemiol 151 (4), 346–357. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Yost K, Perkins C, Cohen R, et al. , 2001. Socioeconomic status and breast cancer incidence in California for different race/ethnic groups. Cancer Causes Control 12 (8), 703–711. [DOI] [PubMed] [Google Scholar]
  17. Yang J, Schupp CW, Harrati A, Ca C, Keegan THM, Gomez SL, 2014. Developing an area-based socioeconomic measure from American Community Survey data. Cancer Prevention Institute of California, Fremont, Calfornia. In. [Google Scholar]
  18. Cheng I, Tseng C, Wu J, Yang J, Conroy SM, Shariff-Marco S, Li L, Hertz A, Gomez SL, Le Marchand L, Whittemore AS, Stram DO, Ritz B, Wu AH, 2019. Association between ambient air pollution and breast cancer risk: The multiethnic cohort study. Int J Cancer 146 (3), 699–711. 10.1002/ijc.32308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Jerrett M, Arain A, Kanaroglou P, Beckerman B, Potoglou D, Sahsuvaroglu T, Morrison J, Giovis C, 2005. A review and evaluation of intraurban air pollution exposure models. J Expo Anal Environ Epidemiol 15 (2), 185–204. [DOI] [PubMed] [Google Scholar]
  20. Laurent O, Hu J, Li L, Kleeman MJ, Bartell SM, Cockburn M, Escobedo L, Wu J, 2016. A Statewide Nested Case-Control Study of Preterm Birth and Air Pollution by Source and Composition: California, 2001–2008. Environ Health Perspect 124 (9), 1479–1486. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Li L, Wu AH, Cheng I, Chen J-C, Wu J, 2017. Spatiotemporal estimation of historical PM2.5 concentrations using PM10, meteorological variables, and spatial effects. Atmos. Environment 166, 182–191. [Google Scholar]
  22. American Lung Association Association, 2019. State of the Air.
  23. Brook RD, Rajagopalan S, Pope CA, Brook JR, Bhatnagar A, Diez-Roux AV, Holguin F, Hong Y, Luepker RV, Mittleman MA, Peters A, Siscovick D, Smith SC, Whitsel L, Kaufman JD, 2010. Particulate matter air pollution and cardiovascular disease: An update to the scientific statement from the American Heart Association. Circulation 121 (21), 2331–2378. [DOI] [PubMed] [Google Scholar]
  24. Doyle JJ, Neugut AI, Jacobson JS, Grann VR, Hershman DL, 2005. Chemotherapy and cardiotoxicity in older breast cancer patients: a population-based study. J Clin Oncol 23 (34), 8597–8605. [DOI] [PubMed] [Google Scholar]
  25. Accordino MK, Neugut AI, Hershman DL, 2014. Cardiac effects of anticancer therapy in the elderly. J Clin Oncol 32 (24), 2654–2661. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Raaschou-Nielsen O, Thorsteinson E, Antonsen S, Holst GJ, Sigsgaard T, Geels C, Frohn LM, Christensen JH, Brandt J, Pedersen CB, Hvidtfeldt UA, 2020. Long-term exposure to air pollution and mortality in the Danish population a nationwide study. EClinicalMedicine 28, 100605. 10.1016/j.eclinm.2020.100605. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Liu Q, Gu X, Deng F, Mu L, Baccarelli AA, Guo X, Wu S, 2019. Ambient particulate air pollution and circulating C-reactive protein level: A systematic review and meta-analysis. Int J Hyg Environ Health 222 (5), 756–764. [DOI] [PubMed] [Google Scholar]
  28. Rider CF, Carlsten C, 2019. Air pollution and DNA methylation: effects of exposure in humans. Clin Epigenetics 11 (1), 131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Rao X, Zhong J, Brook RD, Rajagopalan S, 2018. Effect of Particulate Matter Air Pollution on Cardiovascular Oxidative Stress Pathways. Antioxid Redox Signal 28 (9), 797–818. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kang D, Hamasaki N, 2003. Mitochondrial oxidative stress and mitochondrial DNA. Clin Chem Lab Med 41 (10), 1281–1288. [DOI] [PubMed] [Google Scholar]
  31. Tsai D-H, Riediker M, Berchet A, Paccaud F, Waeber G, Vollenweider P, Bochud M, 2019. Effects of short- and long-term exposures to particulate matter on inflammatory marker levels in the general population. Environ Sci Pollut Res Int 26 (19), 19697–19704. [DOI] [PubMed] [Google Scholar]
  32. Li D, Zhang R, Cui L, Chu C, Zhang H, Sun H, Luo J, Zhou L, Chen L, Cui J, Chen S, Mai B, Chen S, Yu J, Cai Z, Zhang J, Jiang Y, Aschner M, Chen R, Zheng Y, Chen W, 2019. Multiple organ injury in male C57BL/6J mice exposed to ambient particulate matter in a real-ambient PM exposure system in Shijiazhuang. China. Environ Pollut 248, 874–887. [DOI] [PubMed] [Google Scholar]
  33. Zhang Y, Wilson R, Heiss J, Breitling LP, Saum K-U, Schöttker B, Holleczek B, Waldenberger M, Peters A, Brenner H, 2017. DNA methylation signatures in peripheral blood strongly predict all-cause mortality. Nat Commun 8 (1). 10.1038/ncomms14617. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp1
NIHMS1967181-supplement-Supp1.pdf (7.8KB, pdf)
Supp2
NIHMS1967181-supplement-Supp2.docx (59.7KB, docx)

RESOURCES