The Growing Evidence for the Role of Air Pollution in Breast Cancer Development

Summary:

In the article that accompanies this editorial, Wu et al., observed that residential exposure to fine particulate matter was associated with higher breast cancer incidence using prospective data from over 58,000 California women in the Multiethnic Cohort Study. These findings, together with a meta-analysis of findings from cohort studies included in the manuscript, highlight the importance of environmental contributors to breast cancer risk.

Breast cancer remains the most commonly diagnosed cancer among women in the United States (U.S.)¹ and worldwide.^2,3 Although survival after a breast cancer diagnosis is high (91% at 5 years after diagnosis)⁴, there are wide ranging consequences, including treatment side effects, adverse mental health impacts⁵, and high financial toxicity.⁶ Epidemiologic research has identified a number of breast cancer risk factors, including reproductive history (e.g., age at menarche, parity, breastfeeding) ⁴ as well as alcohol use⁴, postmenopausal obesity⁴ and cigarette smoking.⁷ Despite the large number of identified risk factors and the proliferation of interventions to reduce some of these exposures, incidence rates have increased for many groups.⁸ The lack of progress may be attributed in part to the fact that most of these risk factors are not easily modifiable and only modestly contribute to incidence (i.e., relative risk estimates < 2.0).⁴

There is a critical need to identify and quantify the impact of new modifiable breast cancer risk factors, such as environmental chemicals.⁹ However, demonstrating a causal relationship between environmental exposures and breast cancer has been challenging due to difficulties in accurately measuring these ubiquitous exposures, particularly at the low levels at which they may still be harmful.¹⁰ Outdoor air pollution is a particularly promising candidate risk factor given that it contains carcinogenic chemicals¹¹ and endocrine disruptors,¹² which may be particularly relevant for breast cancer given its relationship to hormonal factors. Outdoor air pollution was classified as a Group 1 carcinogen based on epidemiologic evidence for fine particulate matter (PM_2.5) and lung cancer as well as mechanistic and experimental studies,¹¹ and there has been limited investigation into how air pollution is related to other cancer sites.¹³ Even though outdoor air pollution is not amenable to individual-level modifications, policy-driven interventions have been effective in reducing exposure.^14,15

Early epidemiologic studies of air pollution and breast cancer supported an association between breast cancer incidence and traffic-related emissions, such as nitrogen dioxide (NO₂), but studies for PM_2.5 remained inconclusive.¹⁶ This lack of consistency may be reflective of the challenges in exposure assessment and lack of statistical power to detect modest effect estimates. Routine PM_2.5 monitoring by the Environmental Protection Agency started in 1999.^17,18 Some studies relied on air pollution exposure models that estimated past exposure based on more recent monitored concentrations.¹⁹ Additionally, many studies focused on exposure estimated for a single residence (e.g., enrollment address^20–22). These approaches do not allow for the consideration of changes in air pollution concentrations across the different spaces people travel or over time. This may result in substantial exposure misclassification, particularly from residential mobility.^23,24 Indeed, while some early studies of PM_2.5 did not find an association,^25–27 others that did find an association had confidence intervals that were too wide to draw firm conclusions.^21,28 Consequently, two 2021 meta-analyses concluded that there was no evidence of an association.^29,30 However, the tide recently began to turn; large, well-designed studies with sufficient sample size and statistical power to detect a modest association began to accumulate support for a positive relationship.^31–35

In the article that accompanies this editorial, Wu et al.,³⁶ present new findings on the role of air pollutants on the risk of breast cancer in a racially and ethnically diverse cohort of over 58,000 Californian women in the Multiethnic Cohort Study (MEC). This article is an update to a prior analysis, which observed positive but not statistically significant findings.²⁸ With >700 new cases accrued in 8 years of additional follow-up since their previous analysis, Wu et al., used updated air pollution exposure models to estimate residential air pollution concentrations of particulate matter and gaseous pollutants (NO₂, NO_x, ozone, among others) from 1993–2018. Over an average follow-up period of 19 years, over 3,500 incident breast cancer cases were diagnosed.

The key findings of this study include a statistically significant 28% (95% CI: 1.09–1.51) higher incidence of breast cancer for a 10 μg/m³ increase in PM_2.5. The HR was slightly stronger for hormone receptor negative compared to positive tumors (HR=1.26 vs. HR=1.17), albeit with overlapping confidence intervals. For the traffic-related pollutants NO₂ and NOx, 8–10% higher hazard ratios were observed, and although not statistically significant, are similar to previous studies.²⁹

A key strength of this cohort is that this study population is racially, ethnically, and socioeconomically diverse, with over 50% of the cohort having less than a high school education. Most prior analyses in U.S. cohorts, with the exception of the Black Women’s Health Study²⁰, have been composed of predominately white women and those of a higher socioeconomic status.^21,25,35 The MEC study population almost exclusively lives in urban areas, which experience higher concentrations of air pollution. Additionally, racial and ethnic minoritized individuals are more likely to experience higher air pollution due to historic redlining,³⁷ and the inequitable placement of industrial facilities³⁸ and highways.³⁹ Even though there was little evidence of racial or socioeconomic disparities in this study –with the strongest estimates for the PM_2.5 association observed in white women – it is noteworthy that the average concentration of pollutants in this cohort is high across all groups, reflecting the urbanicity of this population. For example, average PM_2.5 levels at baseline were 23.6 μg/m³. In contrast, the average PM_2.5 level was 12.0 μg/m³ in the Nurses Health Study⁴⁰ and 15.6 μg/m³ in the NIH-AARP cohort for a similar time frame (1990–1994).³⁵ In 2023, the mean PM_2.5 concentrations in the U.S. were 8.5 μg/m³.⁴¹

The inclusion of a meta-analysis of previous cohort studies on the topic is an important contribution. This meta-analysis comes at a critical time, as the previous meta-analyses^29,30 did not include the more recent, large, and well-conducted studies that did observe an effect. In the Wu et al. meta-analysis of cohort studies including the new findings from MEC, they estimated a summary hazard ratio of 1.05 (95% CI: 1.00–1.10) per 10 μg/m³ of PM_2.5. This 10 μg/m³ increase in PM_2.5 reflects a large shift in exposure, it is equivalent to going from the 10^th to the 90th percentile in the U.S. in 2000; for 2023, the difference between the 90^th-10^th percentile is about 4 μg/m³.⁴¹ While it is challenging to compare the magnitude of this association to other breast cancer risk factors (e.g., reproductive history, alcohol intake) due to the very different scales, this is similar to the estimates between PM_2.5 exposure and lung cancer risk (9% increase)⁴² and incident stroke (13% increase).⁴³ While the size of these effect estimates is modest, because of the ubiquitous nature of air pollution, the public health burden may be substantial.²⁹ Furthermore, it is entirely plausible that these estimates may be biased towards the null due to the limitations of air pollution exposure assessment.^44,45 Although epidemiologic studies have improved with better exposure models and with incorporating exposure estimates for multiple residences, these studies have largely not considered exposure away from the home (e.g., commuting or at the workplace) or during the hypothesized most susceptible windows such as puberty and pregnancy,⁴⁶ when exposures may have a stronger effect.⁴⁷ Ascertaining exposure during these windows has been challenging due to the timeframe for air pollution monitoring, but future studies will be better poised to estimate air pollutant exposure during these critical periods.

Importantly, the studies on outdoor air pollution and breast cancer incidence have overwhelmingly been conducted in study populations in the U.S., Canada and Europe.^{29,31,32,34,35} Rapid industrialization and urbanization, particularly in middle and low income countries, has resulted in higher air pollution levels, with annual average PM_2.5 concentrations vastly above the World Health Organization guideline of 5 μg/m.^{3, 48–50} Breast cancer is the most common cancer diagnosed among women in China and India.³ Yet, epidemiologic research in these areas with the highest ambient air pollution is limited.^33,51,52 In the one prospective cohort study conducted in China, in >80,000 women living in Beijing who experienced exceptionally high PM_2.5 (5-year average: 78.15 μg/m³), positive linear associations were observed with total PM_2.5, with variation in the association by individual PM_2.5 constituents.⁵¹ Future studies are needed in China and other countries with high ambient air pollution using data from large study populations with strong exposure assessment in order to better characterize the global impact of PM_2.5 on breast cancer.

Despite the declines in U.S. air pollution exposure since the Clean Air Act, PM_2.5 concentrations are no longer decreasing.^15,41 With climate-driven increases in wildfire frequency and intensity, it is estimated that PM_2.5 concentrations will increase ~10% in the next 30 years.⁵³ With this increase in concentrations will likely come a change in PM_2.5 composition – wildfire-related PM_2.5 composition includes higher concentrations of chemicals such as metals and polycyclic aromatic hydrocarbons^54,55, which have been associated with endocrine disruption⁵⁶ and a higher risk of breast cancer.⁵⁷ Future research needs to consider how PM_2.5 composition, which varies geographically, may impact observed associations with breast cancer,^21,22 in order to identify the most important PM_2.5 sources to target for interventions to reduce exposure.

As demonstrated by the meta-analysis in Wu et al., the epidemiologic literature is converging on PM_2.5 as a risk factor for breast cancer. These findings support additional policy-level interventions to reduce outdoor air pollution concentrations, particularly given the projected increases in PM_2.5 due to wildfires. Increased awareness of the relationship between air pollution and breast cancer for both physicians and patients could facilitate more routine capturing of information related to a patient’s residential histories. Although exposure to outdoor air pollution is largely not directly modifiable or treatable, having patients comprehensive residence information can be used to estimate air pollutant exposure concentrations⁵⁸ as well as other residence-based environmental exposures. In doing so, we can pave the way to a better understanding of the environmental contributors to breast cancer etiology.